Nuance Communications, Inc. (NUAN)
December 06, 2012 9:30 am ET
Paul A. Ricci - Chairman and Chief Executive Officer
Mike Thompson - Senior Vice President and General Manager of Mobile Division
Robert Weideman - Senior Vice President and General Manager of Enterprise Division
Janet M. Dillione - Executive Vice President and General Manager of Healthcare Business
Steven G. Chambers - President of Sales & Marketing and Executive Vice President
Thomas L. Beaudoin - Chief Financial Officer and Executive Vice President
John F. Bright - Avondale Partners, LLC, Research Division
Scott Zeller - Needham & Company, LLC, Research Division
Shyam Patil - Raymond James & Associates, Inc., Research Division
Brent Thill - UBS Investment Bank, Research Division
Shaul Eyal - Oppenheimer & Co. Inc., Research Division
Nandan Amladi - Deutsche Bank AG, Research Division
Daniel H. Ives - FBR Capital Markets & Co., Research Division
Jeffrey Van Rhee - Craig-Hallum Capital Group LLC, Research Division
Good morning, everybody, and welcome to Nuance's Fiscal '13 Financial Analyst Day. I'm Kevin Faulkner, Vice President of Investor Relations. We appreciate your interest and time to attend today. We're very excited about our growth opportunity and looking forward to telling you about the exciting things that we have going on. But before we do that, there are a couple of administrative announcements that I wanted to make. First of all, you have the agendas in front of you. I won't go over the agendas, but we are planning both to break in between the enterprise and healthcare sessions and also a question-and-answer session with all the speakers at the end.
I wanted to remind you that today's presentations will include forward-looking statements, estimates, expectations, predictions about the future. These statements are subject to risks and uncertainties that could cause actual results to differ materially, and you should refer to Annual Report on 10-K for the fiscal year ended September 30, 2012, for a detailed list of Risk Factors.
In addition, we will present materials today that include non-GAAP financial measures. We have provided an appendix that's a reconciliation of GAAP to non-GAAP, and you should refer to that appendix. It can be found on the webcast page. It can also be found on the Investor Relations page of our website, and it is included in the soft copy of the materials that will be available today.
Without further adieu, let me introduce our Chairman and CEO, Paul Ricci.
Paul A. Ricci
Good morning, everyone, and welcome. We appreciate your taking the time to [Audio Gap] what complex and multidimensional business. And we really appreciate the opportunity and the additional time to try and delve into it more deeply than we can on typical earnings calls. My role this morning is to make some introductory comments. And then I'm going to turn it over to a series of my colleagues, who are going to go through the business in greater depth. After my introductory comments, Vlad Sejnoha is going, who's our Chief Technology Officer, is going to come up and talk in more depth about our technology than we have done in the past.
Our reason for doing that is that we believe that our technology increasingly provides a source of differentiation for us, and we want you to see the ways in which that differentiation is common across our diverse markets.
Following Vlad, each of our business unit general managers will come up and talk about our 3 primary markets: Mobile & Consumer, our enterprise self-service and healthcare. And then finally, Steve Chambers, who's President of our Sales and Marketing organization, is going to come back and bring that all together with giving you a comprehensive view of our go-to-market strategy and how that strategy is knitted across these diverse markets and how it's evolving over the course of this year. Actually he's not the final speaker. And then lastly, Tom Beaudoin, our Chief Financial Officer, will come up and discuss our financials in greater depth. And then we'll be pleased to take your questions.
So there are really 3 things that I wanted to do. First, I want to expand a bit on the comments from the most recent earnings call. And then I want to focus specifically on what factors are going to drive growth in fiscal '13. And then lastly, I want to return to this point that I have been discussing in recent earnings calls about the migration and the evolution towards intelligent systems.
We think that this evolution towards intelligent systems from a speech recognition to incorporate natural language processing, dialogue management, some forms of reasoning, it provides a foundation for enhanced differentiation for us and provides an opportunity for expanded markets for us, so we're going to talk a lot about it today.
I can say that I think, without hyperbole, that fiscal '12 was the best year in Nuance's history. You're familiar with these financial results, and so I'm not going to go through them. Tom will talk about them in more depth, but it was a year of solid earnings -- solid revenue growth, terrific earnings growth and really extraordinary cash flow generation for us. And we did that while making some important investments in the business in technology, some important investments in the business in sales and marketing. We'll talk more about those later. And we did that while making other important operational strides in the business and making some important strategic moves in the business, all of which we're going to outline for you.
I'd like to start by just a few words about our Mobile and Consumer business. It was truly a momentous year in this business. Our Mobile & Consumer business surpassed $500 million in revenue last year. We had growth of almost 30%. When I was up here a year ago, you were all, I think, quite interested and enthusiastic about our position in the market for Apple devices and iOS, but concerned about our ability to address the Android market. Here we are, a year later, we've cemented a series of agreements with Samsung. We have an entrenched position with them for supplying speech recognition, intelligent system solutions on the Galaxy platform, and we've secured relationships with other important Android manufacturers as well, which Mike will talk about more.
Our view is that we do indeed have a privileged position across most of the leading smartphone platforms in the world today. And that we're going to be able to continue to extend that position based on the differentiation of our technology.
Beyond smartphones, we've talked to you in the past about the position we have in the automotive market, a business that we continue to be very enthusiastic about. We talked to you last year about migrating the automotive business to a connected car business. And in fact, we have now deployed several instances of that. We think what was already an unrivaled position of leadership in the automotive market is now one that's incorporated in the connected car and expanding.
We've extended the business to new markets, including the market for connected televisions. We talked to you last year about that. I think we showed you a demonstration, an early demonstration prototype of Dragon TV. We now ship Dragon TV. We've secured a number of Asian manufacturers. And in fact, I think there are units with our speech that are available for the holiday season. And importantly, we completed -- initiated a very comprehensive strategic relationship with Intel, designed to enable a new market opportunity around the hybrid form factor of laptops and tablets in the Intel Ultrabooks series. That's a relationship that's going to incorporate not just Intel, but a number of Intels and important OEMs. And I think, in fact, there is 1 or 2 OEMs this Christmas season shipping the early versions of that software that we've developed for them.
But we think the examples of the television market and the example of the Intel relationship proved the thesis that we have articulated to you in the past, which is that we're going through a paradigm shift. And that paradigm shift is towards more intelligent systems; virtual assistance; interactive systems that incorporate speech, natural language processing; some amount of knowledge domain and reasoning, and that, that paradigm shift is moving -- that was first articulated in the smartphone market is moving to incorporate platforms and products across the mobile consumer and electronics industry. And we think that opens the opportunity for additional markets that you're going to see as -- over the next couple of years.
Evidence of that is that we now have more than 13,000 developers, who are building applications of one form or another in Nuance's mobile cloud. And in fact, as we mentioned in our recent earnings call, the annualized rate of transactions in that cloud right now at the end of the fourth quarter was about 5 billion.
So we've been bullish about the mobile business in the past. We remain very bullish. I know there's been concern about various competitive forces. We believe strongly in our source of technological differentiation. We believe strongly in the foundation of the relationships we've built with big important customers, such as Ford and BMW and Apple, Samsung. It's a business that we think is not going to be so much challenged by competition, but by in fact the evolution of that business, first to a more cloud-based services and by the extensive customer engagements that we're having to do in that business.
I sat through a review earlier this week of just one customer engagement, where we now have more than 100 people in our Research and Development organization assigned to that engagement. And we have numerous engagements like that. It's created tremendous opportunity for us. At the same time, it's creating a burden of considerable hiring, and we've talked to you about that in the past. For us, that's a very good problem to have because, in fact, we think it defines future growth and expansion of markets for us in that business.
Moving to the Enterprise business, I should start by saying that we had a management change in this business in the middle of last year. Robert Weideman, who's been a long time executive at Nuance, was part of our European operations for a number of years and really built Nuance's imaging business and, more importantly, transformed Nuance's imaging business to the OEM business we have today, working with large MFP partners. Robert, who'd had some experience in the enterprise business previously, has taken over as the General Manager there and has been in that position about 6 months, and he'll have an opportunity in a few minutes to tell you the things that he's doing.
I have said in the past, I believe profoundly in this business. I believe that no one can deliver the quality of solutions in a cloud service for self-service automation other than Nuance can. And we've talked about proof points of that such as USAA and US Air and other examples of really truly leading state-of-the-art applications. And indeed, we did see improved revenue growth this year, after a couple of years, a very tough revenue growth. And in the back half of the year, we saw a very strong bookings in our on-demand business and very strong bookings in delivery of our professional services. So I think our core business is looking better. We surpassed 3,000 deployments. Our cloud is delivering 7.5 billion interactions a year right now.
But as well, we have some new market opportunities that are augmenting our core business. Robert will talk to you about voice biometrics. We're excited about that. We're excited about it in a way it intersects our traditional posted customer care business. We're also excited about the intersection between that and our mobility business.
And Robert will also talk to you about Nina, our mobile virtual assistant solution for enterprises. I've talked about that in recent earnings calls. It's generated extraordinary interest and a very robust pipeline, and I'll let Robert speak about that in more depth.
We think all these things put together are going to allow us to expand our on-demand business and enterprise as we go through the second half of this year and expand the growth rate overall for the enterprise business, based on the complement of the new markets and new product opportunities.
Probably no business saw more change than the healthcare business this year, extremely busy and complex agenda this year. We had to assimilate a couple of acquisitions, importantly the Transcend acquisition. We expanded our on-demand business to exceed 5 billion lines annually. We continue to fuel the growth of our Dragon Medical franchise, been an extraordinary franchise for us. We achieved a resurgence in our radiology business, based on new product offerings that we had under development there. We've now surpassed 450,000 physicians using Nuance's products directly. And we've really positioned ourselves very well to leverage the imperatives and the opportunities for meaningful use around electronic medical records, for the transition to ICD-10, and for the imperatives in clinical documentation improvement, which Janet will talk more about in some detail.
I mentioned in the earnings call recently that we expect that growth of that business -- we expect that business, which is, as you can see, 7 years of extraordinary growth, we expect that business to exceed $1 billion in revenue this year. There was some skepticism about that, but in fact, we think the math supports that. We expect comparable growth rate this year to what we saw last year. That growth rate will be led, again, by Dragon Medical. It will be led by radiology. It will be led by some improvement in our on-demand growth rates. It will be supplemented by our Clinical Language Understanding products. On top of that, we will have the benefit of the new acquisitions in the CTI and computer-aided coding space. And we expect to be probably $50 million of contribution from other small acquisitions that are in various stages of flight right now for the company.
We put all that together, and we will have $1 billion health care IT software and services business as we exit fiscal '13.
The growth of that business is going to be founded in no small part on the very hard work that the company has done over the last year to expand its network of partnerships. These partnerships have been very important for us, and Janet will talk more about them, as we cement our opportunity around meaningful use for electronic medical records. And in particular, we've spoken publicly about the extensive aspects of our partnerships with Cerner and with Epic recently. You'll hear more about those. But there are others, and we have partnerships throughout the electronic medical record space. We have technology partnerships such as UPMC and IBM. I think that will be a continuing theme, and perhaps you'll hear some of that in Steve's discussion in our go-to-market strategy at the end of the day as well.
And finally, in health care, we -- on top of what was a very busy year operationally, we had a vision for how we could transform the use of clinical documentation, creating an end-to-end solution from the early physician-created documentation and issues of quality and the need for improvement around that clinical documentation, through the back end of revenue cycle management and computer-aided coding. You're going hear a lot about that today. We think it opens up an important new opportunity -- market opportunity and revenue opportunity for us. We also think it's going to be transformative in a way clinical documentation is done in health care institutions here.
The components that this new opportunity is founded on include the 2 platforms that we acquired in technology and services, with J.A. Thomas and with Quantim, and includes leveraging the fabric of partners that I referred to a few moments ago and includes the technology foundation that we've been building in our own Research and Development activities inside Nuance, as well as the joint research activity that we've had undergoing for a number of years to create state-of-the-art Clinical Language Understanding technology with IBM.
You're going to hear a lot more about this today. Vlad will give you some more details, and Janet is going to talk about this in depth.
If you look across our markets, there really are 2 important common themes you're going to hear today. They're not new to you, but we're going to go into them in more depth. The first is the importance of deep customer relationships. Nuance has and continues to be the most important trusted partner to deliver global solutions at global scale for our technology worldwide. It's difficult to do. It requires you to bring together a large complement of technologies, not just speech recognition, but natural language processing. Other artificial intelligence technology requires you to be able to deliver them in a large number of languages to be able to engage with significant professional services, technology services and, in some instances, research services, and to do it in a disciplined way at significant scale for a number of years.
And we think that the combination of our ability to project our delivery on a global basis, the combination of the depth of our technology makes us really quite differentiated in this area. I've talked in the past and some today about important engagements with Samsung, with Intel, Cerner, Ford, BMW. We have numerous engagements like this, and we believe that Nuance is viewed in a very distinct way by these companies, by our ability to combine this technology and our delivery capabilities to engage them, so that they themselves can have a differentiated solution that incorporates our 4 technologies.
And the second theme is that we are moving in an evolution to a paradigm shift, which we've -- what I've referred to today and I've referred to in the past. And that paradigm shift is towards a far more robust intelligent system, one that's interactive, one that's speech enabled, one that is most likely mobile, one that is delivered through enhanced cloud services, and one that is customized and somehow differentiated for the partner that we're working with. You can hear a lot about that today. We're very proud of the work we've done there. We're excited about the opportunity this has created. We think that the differentiation there is important. And I think one of the important things that I hope we will leave you with today is the extent of the investments that we have made in that and we're continuing to make in that and the market opportunities that it affords.
So in closing, I just want to say a couple of things. You're going to be seeing an expanded technology set today. A technology set, as I've said several times this morning, is an important source of our differentiation. You're going to be seeing a broadening set of customer solutions and offerings around that technology set. You're going to hear about a broader set of customer engagements that you've heard about in the past. And I believe that all these things will help you understand my comments in recent quarters about my conviction that this really is Nuance's time, that the market has arrived at a place where the set of technologies that we have are unique, that the paradigm shift that's going on in the market leverages those technologies and that the investments we're making now, investments in research and development, enhanced investments in sales people, more investments in demand creation, we do with the conviction that we can accelerate growth as we go through fiscal '13 and into fiscal '14. And with that, I would like to turn it over to our Chief Technology Officer, Vlad Sejnoha.
Thank you very much, Paul. Good morning. I would like to give you an overview of the work we've been doing in Nuance's research organization, and particularly the work we've been doing on evolving our technology core from speech recognition to understanding to what we've been calling intelligent systems that power applications across our businesses in health care, consumer and enterprise. So as Paul mentioned, intelligent systems are systems which interact with users in very rich ways, understand the input and apply increasingly sophisticated reasoning to help us solve high-value tasks. Those tasks can range from easing, accessing important information for consumers, helping doctors understand the interactions between different drugs, carrying out sophisticated transactions in customer care or assigning billing codes to a medical report.
The theme of our work is really nicely captured by this quote from one of the sort of seminal contributors to the field of AI, who basically said that a precondition for machine intelligence is giving machines humanlike senses. We wholeheartedly agree with this statement. What does it mean? It means that if we want intelligent systems to really help us in a meaningful way with problems that we care about, we have to be able to communicate with them in a more flexible, powerful and efficient way. And also that they have to be far more aware of the circumstances and context of the problem that we're trying to solve. So we believe that the technology investments that we're making put us in a very good position in this regard.
Those of you who were with us last year were introduced to our common natural language understanding framework that we used to build applications across the business. I'll just go very briefly highlight some of its salient characteristics. So as I mentioned earlier, we are able to support input from users across a wide range of modalities, certainly voice through speech recognition, but also text, touch, gestures, video. Essentially any other sensor input that tells us a little bit about the user and the outside world can be handled by this framework. Similarly, on the output side, we can communicate with users in a wide variety of ways through text, through videos, through synthetic speech. All of the inputs are subjected to understanding and then on to reasoning.
This is an area where, as Paul mentioned earlier, we've been making substantial investments over the last year, and I highlighted those areas here. These components relate to better, more powerful, more flexible descriptions of the task at hand. And there we are moving from preprogramming, what these tasks are about, towards more dynamic ways of defining goals of an interaction.
And then also more powerful approaches for conducting an interaction towards those goals. These components are informed by an understanding only of the user input, but also of the context, past conversation history, user models and so forth. A lot of metadata goes into improving our understanding of the best next step.
It also uses ever more sophisticated representations of the entities or the concepts that matter in a particular domain. So this is the knowledge representation side, and I'll talk more about that a little bit later.
So I say, is this is NLU or AI? We're really moving squarely into the province of artificial intelligence. And in recognition of this, we have actually set up an AI lab in our Sunnyvale office. Whereas in the past, a lot of artificial intelligence solutions prove to be quite brittle in real life circumstances, the current state-of-the-art in this area has evolved to a point where we're now dealing with algorithms which are far more resilient in the face of ambiguity and uncertainty. We do probabilistic reasoning. We have good fallback methods and, in fact, are able to come up with approximate solutions where exact solutions are no longer possible. So very exciting and more on that later.
But first, I wanted to touch upon some of the developments in our core technology. So speech recognition is really foundational for us. And the ability of the user to interact with intelligent system really starts with us being able to understand what this user is saying. So we are doing a lot of work in this area. And an example of how we are doing is represented by this first chart that shows a comparison of our handset dictation system against a system from a leading competitor. And basically, every point above the dashed line means that the error rate in the Nuance System was lower than the error rate of the competitor. And in aggregate our error rate was half of that of the competitor. Now this is not a stationary field. We continue to improve the system, and the second chart shows how we've been doing over time. And the takeaway here is that we've been on average able to reduce our average error rate on large challenging tasks, about 18% per year through a combination of techniques, which I'll mention in just a second.
So what has changed? Simply put, we're now using orders of magnitude more data between our system, orders of magnitude more complex models to average and synthesize and capture those data, and orders of magnitude more computation to create those models during the training phase.
A particularly interesting recent development was the resurgence of neural networks applied to speech recognition and other pattern-recognition problems. They're called deep belief networks, and we combine those with numerous other modeling techniques in an aggregate composite way to great effect. One really fortuitous fact here is that while it takes immense computation tens of thousands of compute cores in a parallel grid use of IBM's Blue Gene supercomputer.
At one-time, when we deploy these models, we can evaluate them on fairly, at a broad range of realistic devices ranging from the smallest smartphones, as well as the cloud. So this results in very practical and very powerful speech recognition. So while accuracy remains really kind of a holy grail here, we want to improve usability and reduce -- eliminate obstacles to usability no matter how mundane they might be. And this might include simply having to pick up a device, turning it on, flipping through some screens full of application icons, finding the right one, launching it and then interacting with it.
Today, you can already wake up your device by simply giving it a wake-up command and then telling it what you want to do. That already collapses a lot of those steps. But the real goal that we're working on and that will be in the market in not too long, is simply being able to speak to a dormant device that's on your desk without touching it. The device wakes itself up, ascertains that it's you speaking using a voice biometrics, determines whether the input actually seems to be interesting and relevant inputs to it and then wakes itself up progressively more to actually execute it. So you are now talking about simply saying what you want and it happening.
To do that, it's important to do it in a way that is very power conscious. Essentially, this device is listening all the time, and you don't want to drain the battery. And so to that end, we've been working very closely with the principal chip makers, including Intel, on optimizing the low-level silicon architectures precisely for this kind of functionality.
The chip makers understand that we are moving away from the traditional generic computer architecture to one that inherently supports multiple sensors. These are systems that are no longer divorced from the outside world, where the communication happens through a very thin pipe of text, but they're devices are increasingly aware of the circumstances of the user's actions.
So not only are they helping us implement full natural language wake-up in an efficient manner, they're also building in specific support for natural language understanding and speech recognition processing. So these functions now become part of the operating system fabric that other applications can draw on versus competing with for resources.
We further try to be efficient about the use of resources to optimize the user experience by tapping into the cloud transparently to the user. So our natural language understanding framework is flexibly distributable across the cloud in multiple devices. This is very important because increasingly, our users expect a continuity of experience across devices. Whether they're interacting with applications on their phone, in the car or increasingly on their TV. It becomes very important to tie all those experiences back and make them consistent, aggregate learnings and push them out across a myriad of devices. And this is also very important to our large OEM partners, whose stables of devices include many form factors and frequently many operating systems. The aggregation of data is really, really key here. It allows us to continually upgrade models and push them out and ride the performance curve. I've got a particular example of that here. So among the many input modalities we support is tracing on the keyboard, we call it Swiping, where we use predictive algorithms that map the gestures onto character sequences and also perform autocomplete. So this is a very efficient way of entering text in circumstances when you can't or don't want to speak.
As every user uses Swipe, the information goes up into the cloud where it's aggregated and mined and we discover new terms, new language patterns and immediately update the language model for that user. But we do more. We update the language model for other users. So everyone benefits effectively from this crowdsourcing of experience.
You can make this even more refined and apply something called collaborative filtering, where you start aggregating data from users who are similar to you in important ways. You all talk about financial topics in your communication. It's possible that it will be a financial language model that'd be generated for you and your like-minded users. This is something we do across all the modalities, not just swiping, but also speech recognition NLU. So the core is getting very good and very robust, and it's becoming easier to take one of our cloud APIs and simply plug them into an application and achieve good results. And so we have offered up our APIs, now in 38 languages, to the developer, an ISV community.
And as Paul mentioned, the uptake has been really impressive. Applications are being released. This example is just a recent one of an electronic health record system from Iconix that uses our speech recognition input. This is true across our businesses, so there are applications in the consumer space, the health -- in healthcare space, as well as well as enterprise.
And again, this is a very efficient way for us to learn about new application use cases without necessarily having to build every prototype application ourselves. We see all the data that's coming through. Really invaluable.
One application category that's really taken off in the last year or so has been that of the natural language understanding application or more precisely, the virtual assistant. So obviously, there was Siri, a huge visibility; our own DragonGo! Dragon Mobile Assistant; now there are about 60 and counting, including Google. But comparing these is not straightforward. When you look at speech recognition and strip away a lot of factors, so the important one that remains is, how accurate is it? Does it really recognize my speech with sufficient accuracy? When it comes to a virtual assistant or a natural language understanding application, things are a little bit more complicated. Sure, the accuracy of the NLU matters, but there are always other questions that you want to answer. One is, how robust is the system to inputs, which were not anticipated necessarily by the designers? Does it fall apart? Does it behave intelligently? How efficient is it? Is it -- if it attempts to achieve some goal. Does it take many steps? Is it very, very quick? How broad is it? How many functions does it support? How deep is its understanding within any one of those areas? How portable is it? Does it have a persona? And if so, is the persona a good one, or is it annoying? So how do we go about tackling this problem?
So let's start with NLU. So how do you build a powerful NLU? There are systems out there which depend nearly exclusively on machine learning from big data, where you learn patterns and capture those patterns through observations. On the other side, there are systems which depend nearly exclusively on the handcrafting of rules. We're convinced at this point is ample evidence that neither a stand-alone approach is sufficient, that you really need a judicious synthesis of these 2 to build a state-of-the-art system. We start out with a linguistic theory, a framework, and then you populate it through observation and to take advantage of massive data.
There are ample examples how these extreme approaches fail. I've just got one here. So all 3 major search engines rely very heavily on data to map queries onto URLs. It's really easy to find examples where the lack of a linguistic framework causes problem. So all 3 handle this question incorrectly, "Who did Business Objects buy?" The top results here -- the search provider is anonymized here -- are wrong because you lack a linguistic framework that tells them that the subject of the question is "Business Objects," it's not the object. We don't want to know about why did SAP buy Business Objects, but vice versa, who did Business Objects buy. So statistical training but within the context of linguistic framework is key.
Similarly, it is important to find an efficient way of capturing the content, the domain-specific vertical information about a new area that the NLU has to understand. And this is the area of ontology, which are structures, which layout the important concepts, their characteristics, their attributes, how they're related to each other, and what actions you can perform upon them, what is reasonable to do here. These are essential for sophisticated NLU, and they're also essential for the efficient creation of applications, additional applications, within a given domain because they are reusable. So we're doing a lot of work in building these out across vertical domains. And just as one example here, our LinKBase medical ontology has 1 million medical concepts and 3 million relationships. And this is really instrumental to our work in healthcare NLU.
So our Clinical Language Understanding is a great example of the application of linguistic structures, statistical learning and the use of an ontology. So as you probably all well know, doctors prefer to document patient encounters through narrative dictation. The results are very information-rich, but they're difficult for others to consume, to read through. And further, it's already mandated that the essential information contained in these reports be entered exclusively into electronic health records, databases. Doing so is very laborious. Doctors don't like to do it, so our system automates the extraction of these facts. And the last time we spoke, we were in the middle of this curse, and we've made a lot of progress since. We're now at the 93% accuracy level, which is approaching human performance on this fact-extraction task. So this is really kind of a very important event for us.
But the extraction of facts is just the first stage. Our ambition is much greater than this. And we want to automate the whole clinical documentation workflow. So you have the facts. We are able to immediately check whether the information the doctor provided is sufficient, whether it's complete, whether it adheres to documentation standards. We can do that automatically. We can also automatically come back to the doctor with a clarifying question, prompt the doctor to provide additional information or disambiguate something that is ambiguous. The resulting document is of a higher quality immediately, which is really important.
And then we can automatically start assigning billing codes, ICD-9 and ICD-10, without which the hospital or the doctor cannot be reimbursed. So a huge commercial opportunity and a very exciting technology area for us. And Janet will talk about this in more depth later.
Earlier, I talked about the potential of applying linguistic processing to reordering the results from searches. But you can really do more for a lot of queries, particularly navigational ones, where the user knows where he or she wants to end up, you can use this NLU framework to completely bypass the intermediate screen of links. So search providers begin to experiment with providing actual answers were a subset of user queries, but none of them are willing or able to yet to really just take you to where you ultimately want to go. So contrast the traditional navigation approach to simply resolving a query directly and taking you to where you want to go. Of course, you would never be able to get a table at the French Laundry the same night, but -- you get the point.
We can further improve the accuracy of query resolution by considering the conversation context of the past history. This is an application, a solution we deployed based on our natural language understanding framework for a Chinese bank to automate the answering of questions their customer sends to them via text messages about their financial instruments. And this entailed about 60,000 topics.
So in the past, they had operators who responded to these questions. We now apply our NLU. Now if you can see from this example, the user asks multiple questions, we carry the context along to actually resolve it. So here, we used the pronoun "it" to understand what is being asked about. And we do that quite extensively. The chart shows the performance while we were training up the system using data and improving the models when we deployed. It's now about 90% accuracy or automation. The customer is very happy with this.
Natural Language Understanding, of the sort that we are building is also transforming customer care. For one, users or customers dealing with companies are bringing in expectations they're gaining from the use of consumer-oriented virtual assistants like Siri. The second factor that's driving the transformation is that customer care applications are very transactional, and it's proven very challenging to represent those complex transactions either as IVR voice menus or even a sequence of visual menus in mobile applications, and either one is successful in that respect.
Natural Language Understanding here has the potential of allowing users to drill through or even bypass these structures and get directly to what they want. And then further through conversational dialog, that we call mixed-initiative dialog, users can provide information proactively to the system and not necessarily follow a scripted and prompted transaction. This has been met with great enthusiasm by the customer care industry, and we have packaged our Natural Language Understanding framework into an SDK we call Nina, that is customizable. So it becomes brand-friendly. And this is very important to these customers because this is their face to the public. And Robert Weideman will be talking about this at great length.
So we are going even further our ambitions in the kind of conversations we want to support are even greater than mixed-initiative dialog. And we are working in our AI lab on something we're calling advanced collaborative dialog, which starts from the assumption that the user and the system have complementary information. They're both going to collaborate to make this information bring to bear on that particular task. And this involves something called the theory of dynamic intention structures. And the example here shows the user interacting with a TV virtual assistant. And you can see that the level of spontaneous questions here is much higher. These questions require the system to dynamically adjust its goals and use reasoning to bring in information that's relevant to advance the dialog. It's also using the user history and a model of the user preferences. So this is very powerful.
So we have a framework, which is very powerful. It offers a lot of functionality, and it really allows us to rebuild the conventional visual user interface from scratch. What do you do now that you understand that through natural language. You can access and manipulate objects and information that aren't even necessarily visible in front of you? How do you rebuild the UI? And there are a lot of design choices, and I listed just a few here. Should you build something that is like a homunculus, a persona, a virtual assistant that has a personality that is very expressive and mediates between you and the system? Or should you build something that's much more blended, something that's more ambient intelligence? Where you do see icons, but you understand you can talk to them, and you can talk, sort of beyond the screen. When you talk to other applications, do they need to be reimplemented in your assistant, or can you talk to them in their native state? These are all really important design choices. And our R&D organization, along with our user experience design group, work with our customers, the customers that Paul mentioned earlier, to build specific instantiations of customers' vision that differentiate them and that achieve the kind of functionality and efficiency that they're hoping for. And this is really important. This is where they find the flexibility of our platform very compelling.
So this is an ambitious agenda, and I just want to give you a quick feel for the kind of investment that we've been making in this regard. So our R&D is about 1,600 people worldwide. Of those, sometimes rough to -- difficult to characterize exactly, but about 400 are researchers, who do fundamental work in advancing core algorithms and techniques and modeling. We also work with 65 people from IBM Research. Of the 400, about 150 are focusing squarely on Natural Language Understanding, and part of that is this AI lab in Sunnyvale. And that's staffed with some of the world's leaders in this area. Then, we have 1,900 professional services people around the world that help us instantiate this -- the technology framework for particular customers.
So in sum, we're developing a very powerful new framework for the next generation of intelligent systems. It's meant to be general, but also meant to be focused on vertical areas and power applications across our businesses: consumer, enterprise and health care. And in most cases, it's done with -- in collaboration jointly with customers. It's a distributive platform, it's an intelligent platform, it's a platform that increasingly understands how to utilize all the customized sensors and specialized architectures on devices. It's one that's distributed, uses the cloud effectively, aggregates data, pushes out learning. It is a usable one, it aims for a consistent experience across form factors and endpoints that a user might be using. And most importantly, it is brand-friendly, which is hugely important to our customers.
So our platform has all those characteristics, and we think this really advances our competitive differentiation. Thank you.
So now let me introduce Mike Thompson, who's the General Manager of the Mobile Division.
Testing, testing. Can you hear me? Thanks, Vlad. That was a great session. Actually, it's -- there's a lot of density in that, but in the real world, everything Vlad just said matters a lot. These are very sophisticated systems.
So I'm Mike. I run the Mobile Division, and I wanted to talk to you guys a little bit about our solutions, our customers and our growth outlook in the mobile business.
We had a good year in 2012, and I think there's a lot to talk about. We started out with this vision to humanize the interface to digital devices. And at first, it was little, tiny feature phones back in 2004, 2005, and it was kind of a crazy idea at the time, or a lot of people thought we were crazy. It's absolutely taken off. And as smartphones took off not long after that, it became clear to us that it wasn't just about interacting with the device right there for a moment in time, but actually connecting to the mobile web, the connected world that we're all now experiencing. And I think the single most important point that I can make, to start, and Paul touched on it a bit, is this vision, this mission, has absolutely become reality. We are at the epicenter of a revolution. Almost all sophisticated mobile devices today are shipping with some kind of capabilities in the areas where Nuance participates. And the outlook for that is only increasing. So we got a lot to talk about today.
The Dragon family. The Mobile and Consumer Division is built on the Dragon family. And based on our outlook of where this market was going, we've been building a pretty sophisticated machine. Vlad talked a lot about the underpinnings of that machine, but this chart reflects the overall breadth and diversity of what we're doing. I will drill down into this as we go, but there are 5 major product families.
Dragon Mobile includes all of our mobile phone and tablet solutions. It's our most mature market. It's absolutely been rocking for quite some time. We'll talk about the different solutions in this space. But there are embedded technologies, things we put on the device and we charge a royalty for. Increasingly, those devices are connected on a permanent basis to our cloud-based service. The mobile business has been very, very important for us. Vlad mentioned Swype. Our keyboard technologies are a big, big part of this, and we'll talk about it more as we get into it.
Dragon TV. There's this megatrend going on that started with smartphones but has extended across this entire slide, and that is smart devices. Smart televisions are now starting to shift in high volume. They're the fastest-growing part of the television market, and we built Dragon TV to make consumers' interactions in the living room or on gaming systems that much easier, that much more natural. Again, they are embedded technologies in televisions, plus those televisions are now connected to the cloud.
On desktop, the original Dragon NaturallySpeaking, the foundational product we started with when we built this company, has been historically an aftermarket product. You bought it in a box down at some store, and then you installed it on your computer. Well, that's changing dramatically. The desktop, laptop, Ultrabook space has converged with mobile phones and tablets. And the entire Dragon desktop portfolio is being optimized and driven towards an OEM business. This is a major shift for us. It'll include embedded technologies and connected technologies.
And then the car, Dragon Drive! The use case in the car is quite an interesting thing. Safety in the car has driven our Automotive business. We have a phenomenal Automotive business. And we actually ship on quite a few cars. But the fastest-growing of the car segment is connected cars or smart cars. Cars are just big devices that want to have relevant information piped into them for drivers so that the driver experience is safer, better, more productive and more entertaining. So we built Dragon Drive! with embedded technologies to sit in the car and connected capabilities in the cloud.
And lastly, voice-to-text -- or voicemail-to-text. This is around the huge voicemail-to-text market. I said to some of you who were here last year, no one likes voicemail anymore. Voicemail is an outdated technology. And we have been dragging that into the modern era with our voicemail-to-text conversion systems and visual voicemail clients. That is changing rapidly as well. It includes embedded technologies and connected cloud-based capabilities.
So the whole portfolio is quite diverse. It's aligned with the megatrends around digital devices, and you'll continue to see this expand as we build out and grow over the coming year.
I wanted to start early on with why Nuance wins. Mobile is a very fast-moving business, and we at Nuance spend a lot of time on strategy. And you've heard some of the key themes that Paul talked about. But we spend a lot of time on this. And these are the pillars of why the Mobile group wins. Vlad highlighted the innovation. We have so much dedication to the products that we're building and so much focus on them, that continuing to lead across that portfolio of things is the #1 mission, absolutely. And we do it across platforms, form factors, operating systems. There is no company in the world with the diversity of deployments that we have when you map it against all of these different ways we get to market. This third mark -- this third bullet is very, very important. Everything we do is configurable, modular and optimized for device makers and carmakers. In this case, device makers, I mean that very broadly, all kinds of device makers.
We're designed to build solutions for them so that they can differentiate their products, and that's what's been fueling our growth. They can, in their own competitive markets, go build a solution around our solution to win themselves in the marketplace. There's more than that. Paul mentioned 100 people on a given project. That's been the biggest explosion of activity around the operation side of the mobile business. The appetite from leading companies to engage with us to build things and expand things and grow things around these solutions in this epicenter of activity has been dramatic. I have multiple teams of 100 people working on multiple projects on multiple continents. And it's been quite an interesting ride to participate in some of the most secretive device designs going on with the most important companies, including clean rooms and locks and special privileges and a lot of other things. But it's critical to our strategy.
These last 2 bullets are important too. Nuance is Switzerland in a way. We don't have a content initiative, we don't own a social network, we're not a search company, we are open. And to device makers who sell products all around the world, that makes a big difference. We can send search to Google, we can send it to Bing, we can send social networking to Facebook, we can send it to Google+. It doesn't matter. We can give options. And to device makers, that makes a very big deal.
That leaves the last point. These device makers realize in the connected world that there's tremendous value on the data they're generating and the knowledge of what people do. And we collaborate with them and share information about those things. And that is providing new value and new opportunities for these device makers that they haven't necessarily had visibility to as just device makers. These 6 things have been driving us and have been working over the last year.
These are our financial highlights. For the last 4 years, we've grown between 25% and 30% for every year for the last 4 years. It's been quite a run. But what's worth calling out here is that back in 2009, we had no connected revenue at all. We just sold small bits of software on devices. And when it shipped, we collected a royalty. Over this time, we've seen significant growth, but we've also held our profit as we built up this connected business. And that's been quite an interesting run. The landscape of what we offer to customers is bigger, and it's deeper, and it's longer term. So that's a big dynamic that's going on underneath these solid financial results.
Financial numbers are one thing. I think it's worth showing just some other numbers that represent the diversity of those financial numbers. These are some numbers, facts and figures about the Mobile & Consumer business. We launched Dragon TV in January of this year. We have now shipped 750,000 voice-enabled smart televisions. The best televisions at Costco or any other retail store, you go to the high end ones, like Samsung Smart TV, those have voice recognition; LG, they have voice recognition. That market is going to continue to grow. Smart connected TV is not going to slowdown.
For the first time, we sold 1 million copies of Dragon NaturallySpeaking, our legacy DNS product. The major product launch this year, just a few months ago, a huge product. People love the product. And the sophistication of that solution, it's the best in the world for a fully functioning desktop solution. 20 million voice-enabled cars shipped last year, that's strong penetration. You're going to see the continued expansion of that, and I'll talk about it on our growth factors.
250 million voice-enabled phones. The smartphone curve is destroying feature phone. And the opportunity for bigger, more sophisticated solutions on smartphones is so clear in front of us.
Keyboards and touch screens. Buttons are dead. You won't be seeing buttons much anymore. There's a place for them on occasion. But buttons break, they're expensive, and they're fixed. The thing about touchscreens is they're dynamic, they don't break, and you can do so many different things on it. So our touch gesture capabilities around Swype see tremendous potential for growth, not only in mobile phones, but in the broad array of tablets and everything in between, where a fixed button-based keyboard is not possible.
And then 1 billion voicemails transcribed. I think I said last year that people hate voicemail. They really do. We're trying to wipe it out and allow people to interact with their voicemail systems in a digital way, in a modern way. But 1 billion, that's a big number. So those are some facts underneath the financial numbers.
This is a picture of the cloud. Dragon cloud traffic was, as I said, 0 in 2009, when we launched our first Dragon dictation downloadable application. We have grown to and done 5 billion cloud-based transactions since Q1 2010. And we expect that this number will, based on the curve, actually grow even faster. But the main point I wanted to highlight here is the diversity of the mix driving this volume. Again, I think we are the only company in the world with the diversity of the following things: we process traffic from IOS, Android, RIM, Symbian, many, many different flavors of Android, so operating systems underneath this huge form factors. It started with smartphones, they took off; then tablets, they took off; then desktop, then cars, then TV. When we just -- when we light up a new segment, it's like watching the growth curve just take off and go from 0 transactions on TV at the start of this year to a number now, those 750,000 shipped TVs are now hitting this cloud-based system. All right? Further, that we got developers on here. Developers are building all kinds of applications, and then we've got a wide range of customers. So this is a very, very diverse, broad, deep representation of cloud-based activity.
Now a lot of companies draw up a logo slide; it's pretty easy to make a logo slide. But this is an important logo slide. There's some really important companies up here, and I wanted to highlight a couple of things. First, let's start with Samsung. Shipping sophisticated speech capabilities and keyboard capabilities on the Note II, the GALAXY III and the high-end, best-selling Connected Smart TV, all 3 from Nuance. And that's indicative of the largest consumer electronics company in the world's commitment to these solutions and where they're headed. I just got back from Korea, I was there last week and I can tell you that this is serious. The potential for where we're headed, not just with Samsung, but with this marketplace is just incredible. All right?
Amazon. New to the device-maker space, we don't talk about our customers in great detail, but they just announced a new Kindle yesterday and go read through the keyboard highlights that are referenced on the site, phenomenal swype-based capabilities from the new Kindle. All right?
And Apple, they're doing some amazing things around voice. And I think the commitment you've heard from Tim Cook is that you're going to see incredible amounts of expansion around Siri.
And then shift gears to the Ford, GM, BMW, that's the automotive line. I didn't have enough space to put the automotive customers up here. The automotive community is so committed to voice recognition because its relevant to the driver, who's got his hands on the wheel. The automotive space is absolutely rocking right now, and it has great potential in front of us.
I mentioned the nature of this evolution of not just selling an aftermarket product, the best-in-class aftermarket product, but selling a variety of products as part of the computer, the desktop computer or the laptop computer, from the factory. That's where you see Intel and Dell, our Dragon Assistant product now shipping with Dell. And of course, our carrier customers who are engaged with the voice-to-text solutions. When you have 1 billion voice-to-text conversions, that's a big number that affects these operators. It's a big commitment that these operators are committing to and generating value around in their services, around their voicemail systems.
All right, so that's a little bit of 2012. We've built this machine. This is illustrative of many of the products that we have. And I think I want to take through a few of these just so you get a feel for them. DragonGo! we launched this summer and a half ago. It was the first Natural Language Understanding application, powerful Natural Language Understanding application in the marketplace. It has set the pace for a lot of what you're seeing now commercially. Dragon TV, I've talked about a lot. This is not going to slow down.
Competitiveness in TV is just amazing. And when the market leaders already have the solution and plans to expand it, you're going to see more. We announced Dragon ID this year. That's voice biometrics, the ability to ID somebody for a simple solution as unlocking your phone without looking at it, without having to touch it, or more sophisticated solutions like commerce. Great exciting things in front around commerce and mobile phones.
Talked about Dragon Drive!. It's our automotive solution, unbelievable solution. Connected capabilities are the big new thing. Visual Voicemail. Why have a voicemail system at all? Just get it presented and dictate right back to it. Our Voicemail to Text solution continues to soar. Dragon Dictation and Dragon Search, those of you who have followed us for a while, we launched these back in early -- late 2009, beginning of 2010. We just launched in, I think, our 38th country, Dragon Dictation and Dragon Search. 38 countries over that period of time, we reached, I believe, #1 status on the app store in 35 of those 38 countries. That's one of the best apps ever launched on the app store ever, Dragon Dictation. And that shows the appetite, not just of U.S. citizens but of the world population for these kinds of capabilities. So there's much, much more to come, and it's going to be global.
I mentioned Swype to tie a point Vlad made. When you buy a phone, you interact with it all the time. You have a keyboard, keyboard gets smart about you. You add words to it and stuff. But then, it's time to buy a new phone or maybe you want to buy a tablet. You get a new keyboard, and then there's no way to get all that smart stuff that was in your original phone into this new keyboard. Well, the new version of Swype, the living, learning keyboard, quite different. Keyboard comes along with you. You can pick it. You can use it, and it can follow you from device to device, different brands.
The Genius Button got smarter. They incorporated Natural Language Understanding. The Developer Program took off. Dragon Assistant, this is what you'll see from Dell, the Intel-Nuance combination, driving through Dell; Dragon Dictation for the Mac; the NaturallySpeaking for the desktop. We had a really, really strong quarter last year with the launch -- just last quarter with the launch of these products. Again, far and away, the most sophisticated solution for complex capabilities on a desktop or a laptop. We launched Dragon Assistant, powerful new solution, highlighting some of our new capabilities. All of this is driving this, and those are the growth drivers.
If you look at each of the business segments, what are the major growth drivers? In the handset and tablet space, you're going to see the continued expansion of personal assistants. Personal assistants are here to stay. They're going to be very unique. They're going to be customized. They're going to be branded. There will be choices. They will get smarter. They will become advisers. They will connect to many, many things. The personal assistant marketplace is going to explode in the whole category of mobile, tablet and related form factors.
As I said, about keyboards, touch is here to stay. You're going to see touch on many, many devices, and there's tremendous expansion potential as smartphones continue to erode feature phones and expand.
On TV, let's put some numbers to this. Less than 1% of TVs today ship with speech of any kind. Embedded capabilities and connected capabilities will ship and penetrate this market as it grows for the coming years.
Similarly, on desktop, our historical business has been fairly low penetration. It's an aftermarket product, Dragon NaturallySpeaking. You're going to see this accelerate as we sell embedded-based capabilities jointly with the manufacturers of computers, desktop computers that ship from the factory.
On automotive, penetration of their embedded capabilities, the software we put on the car and we never hear from again, it's actually pretty good, 30% or more. Connected capabilities, that is speech-enabled capabilities to the cloud, 0% penetration just 6 months ago. As Paul said, Audi and BMW now connected voice-enabled capabilities in the car, tremendous headroom to grow into that business.
And finally, our Voice-to-Text business, there are many, many Voice-to-Text messages -- voicemails left around the world everyday. We have very low penetration relative to the size of that massive market, and we are working very hard to erase that market entirely by converting voicemail into a mega messaging base digital system. So those are the growth drivers for the mobile business. It's what been driving us over the last 4 years and will keep us going.
This is the theme for 2013. A world of personal assistants, and there are some important elements on this slide. Vlad talked about technology and products, all the way to the right. The breadth and depth of our products is really something. There's the design services. We take those products and we combine them in unique ways for our device and auto partners. There's the cloud. We are generating massive amounts of data. It's making us very smart. This thing about intelligence is based on data. It's hard to capture that if you don't have the volume. You can't get smarter if you don't have the traffic, and we have such a broad and unique amount of traffic. We're becoming very smart. We're getting very smart in how we build our products. That's one dimension. We're also becoming very smart about what people do and what they like, and that's creating new business opportunities for us, where they want to go. So this idea of intelligence, smarter systems, which has been central to this discussion, is foundational to 2013 for mobile.
I want to give you a sneak peek of a special project called Wintermute. It's a codename. We just announced this to some of our salespeople last month. And Wintermute is a pretty exciting concept. Like the example I gave with the keyboard, if you have a personal assistant that you've become familiar with on your device and you interact with on a regular basis and it knows you, in fact, it's more of an adviser and you talk to it in a certain way and you've got familiarity and it's something you care about. As you move through the rest of your life, wouldn't it be great if you could take that with you? When you're onto the car, you could be able to interact with your car. When you're interacting with your TV, that same personal assistant would be there.
You're going to see a video here where we'll demonstrate that. Sean Brown, one of our Product Managers, will take you through a quick video here. And he's going to show you 2 things, 2 different personalities, which is indicative of there's going to be many personalities, 2 different personalities; and he's going to show you a personal assistant moving between devices.
So that's Sean. A couple of things went on there. First of all, he interacted with one device and as he moved to the next device, he could talk in a very open, natural way, and that device was informed and intelligent about what was going on before. And this is illustrative of where we're going. It's like a human interaction. There's intelligence there, you can bring it along. This is where we are headed next. We have built this solution. I'm excited to say you can go look at some of the capabilities in the demo rooms outside to get a taste of where this is going. But just like with iCloud or any web-based services, this is where Nuance is taking the next generation, not only within one OEM, who maybe builds phones, tablets, computers, cameras; but across devices, so beyond brands. And that's a very, very unique, powerful position to be in.
Occasionally, I get asked about Google. What's Google doing in speech? How does it affect you? I thought I would talk about that directly here. Google is a big important software company. They have incredible search, and they have a powerful advertising engine. But our business is growing and thriving despite Google. And we're winning for many, many reasons. We're not just focused on search. We're not just focused on advertising. That's not our primary business. That's not what we have at stake. We've got a much bigger thing going on. The breadth of what we're doing and our position and strategy in the marketplace is totally unique. We are optimized for device makers and car makers. We do what they want. We give them tools to create unique things. And that's making a difference in their choices to go with Nuance for very, very large projects for the long term. We customize. We also don't have any of these conflicts. Google has a lot of conflicts, and I'm not going to be the judge of all those conflicts. They've got a lot of conflicts. Our Switzerland position with these device makers is quite unique. Plus, we share with the device makers. That's a very, very big deal. We share data and insights about how people use these systems, how often they use them and what it looks like, what's important, where's their new value. Many of these device makers are seeing an opening to new value through this window, and that is something Google would be very hesitant to give up.
Lastly, it's important to highlight, we're way more than a speech recognition company. If you look at this equation of Natural Language Understanding and reasoning times conversational systems and what a conversational engine does, times the speech recognition, times expressive human interactions around text to speech, times the form factors, times the platforms, times the use cases, times the languages. Google is not doing this. They are focused on saving their search business, which has stalled on the desktop and moving it to mobile. That's a very specific thing. So it's a big company. It's important, but it doesn't get in our way. These are the factors driving our growth.
So I'll close with this picture. The machine is largely built. We are growing in each one of these segments, and these segments are growing. We sell embedded capabilities, connected capabilities. The sophistication of both is going up dramatically, and the megatrends in mobile are absolutely blowing in our sails. So it's an exciting time for the Mobile & Consumer business, and I look forward to talking to you guys later on.
Now I get to introduce Robert Weideman, General Manager of the Enterprise Business, and then we'll take a break after that. Thank you, Robert.
Thank you, Mike. Very, very exciting stuff, Mike, and I'm pleased to be representing the Nuance Enterprise division. I'd like to first tell you a little bit about the Enterprise division in case anybody is new to that, provide a little look at our fiscal year '12 priorities and results, but more importantly, show you the context of our growth strategy and how it relates to a lot of what you've already heard and how it really applies to transforming the customer service experience as we look ahead.
The Nuance Enterprise division is simply the best at delivering automated customer service solutions that not only enhance the user experience, allow people to get access to information through automated cell service but also allow organizations to reduce their operating costs. And it's not just limited to the context, and we're seeing our solutions actually move into areas, mobile that you'll see a lot today and moving them to the web, that really allow people to deliver customer service experiences that are productive and efficient and convenient across multiple channels and do that in a way that their customers can get access to those solutions anytime, anywhere, and most importantly have the same elegant, nice user experience no matter how they access that customer service.
We have a really significant advantage in our core technologies. And Vlad spoke a lot about the accuracy advantages of, for example, in our ASR engines. It just really, really matters for customers that when they say something, the system understands what they say. And accuracy is really job 1 to get started with a positive experience. You don't want to frustrate customers with speech recognition that doesn't understand what you say. And in fact, you can see through the demo that we have outside, we can do this actually in pretty noisy environments, and so it's a very, very strong engine that we've developed in speech recognition. Then, when a system talks back to you, you don't want a robotic voice talking back to you, and we've got really the best text-to-speech technology and deployments of places like US Air, where you can actually see that it's really seamless with the recorded prompts and actually sounds like a natural part of the user experience.
More recently, we've really been expanding our investments in Natural Language Understanding, and this means that we not only know what you said through speech recognition, but what you'd like and what you really meant and how do we deliver answers based on that. And that's becoming a very important part of our portfolio and a very important part of customer experiences through the contact center, but also through the different channels that people are engaging with. Because we service very large organizations and in fact, service smaller organizations throughout the world, we've got the broadest language coverage in the marketplace. And so our ability to attack market opportunities and growth potentials in emerging markets like South America and the APAC region is another advantage that we have.
And we've overlaid all of this with a deep investment in voice biometrics. And so we not only know what you said. We not only know what you want, but now we know who you are when you said it. And that combination, this portfolio that we have is unique in the industry and being able to deploy solutions that really automate the customer experience. And we deliver that through on-premise solutions using our global partners that include Avaya, Cisco, Huawei and Genesys but also through regional partners. So we got a very strong channel partnership organization around the world and through, of course, our on-demand cloud-hosted solution, which allows us to actually bring together this full portfolio of technologies into really the industry-leading solutions for automated customer service.
And it proves out in the numbers. If you take a look at how we're using our frequently used, we've got a really scalable solution. You don't trust somebody with your customer service experience if you're not able to actually handle large volumes, very high predictability levels, uptime all the time. And you look at the numbers, 12 billion calls a year handled through our systems and that, I think is, a conservative number because we can't track everything that goes through our channel partners. We can absolutely track the interactions on our on-demand service, 7.5 billion interactions with customers with high levels of customer satisfaction. We've deployed, as Paul said, over 3,000 enterprises already, 8,000 speech deployments, a fantastic business that we have here. And we've mentioned our strong position in voice biometrics. We've got a 70% share of voice profiles, voice enrollments that are in the field today. Just to give you a context, that's 23 million voiceprints. That's 23 million people that are engaged with our voice biometrics systems, a very strong position that we have there in a market that is really at its starting point in terms of value to the market and penetration into the industry.
And all of that results in organizations that do trust us. The leading organizations in the world have selected us to be their first front door to their customers. When somebody calls, Nuance picks up the phone and says, "How can I help you?" And when you take a look at the industries that we participate in, they're the ones you would expect, but they're pretty broad. It's the financial industries, the insurance industry, it's the government organizations. When you think of interfacing with service through your telco organization, those solutions are delivered by Nuance. And in fact, if you look at what we've had recently, 2 large voice biometric solutions, 1 through the Australian government through the Department of Human Services that are going to use voice biometrics to make sure that people are still alive and ready for their benefits; and also for Barclays that is using voice biometrics in their high-wealth desk. They're actually trimming. It used to take them up to 6 minutes to identify if a person who was they were -- said they were when they called in on the phone. That's now brought down to 30 seconds, and it's done in a passive way. It's not asking a high-wealth individual, what's your name? What's your social security number, when were you born? It's allowing them to speak to the agent, identifying who they are. The agent gets a pop-up on the screen, a seamless conversation they're having with that individual. It allows them to maintain the relationship with the customer without being annoying and still find out and verify who the person was.
So very powerful solutions, again, deployed at organizations that really, really care about their customers. And we form, as Paul said, deep relationships with these organizations through a professional services team that really is simply the best, the best at understanding the problem through business consulting. The best at designing solutions that will deliver high levels of customer satisfaction and automation and actually deploying those solutions and maintaining those solutions for our customers in ways that meet the business objectives of our customers.
That's resulted in a really strong year for us last year and one where I think and what I'll show you really primes us for growth as we look ahead. We actually achieved 12% revenue growth, 7% organic growth in the past year and really focused on our operating metrics in order to deliver a higher level of profitability for the organization. And we did that through a series of strategic and operational focus that we had in the year. We really do believe that the industry is accelerating towards cloud-based services. And particularly, in the latter part of the year, we saw very strong bookings for our on-demand hosted solution. And in fact, you'll see things that we did in the business that I think is going to accelerate that.
Our pro services organization, we actually added capabilities, particularly in Europe, that really paid off in growing our solutions business there for both on-premise or for the on-premise and channel partner programs. But we also expanded our market opportunities. There's voice biometrics investment, and our Nina mobile solutions have entered us into areas of solutions that are providing synergy into our traditional markets in the contact center, but also allowing us to enter new market opportunities such as mobile customer service that we think are going to prove fertile ground for us to grow as we look ahead. And as I mentioned, we really are focused in managing this business much more tightly, so as we invest in those focused areas to grow, we're doing it in a way that we can actually maintain and actually improve our financial performance as we look ahead.
If you take a step back and say, so what's going on in the world of customer service? This is really a transformational period for customer service. Back when interactive voice response systems, touch tone systems were invented, that was a revolutionary change in the industry, providing a new technology foundation for people to interact with customers in a more efficient way, efficient for the customer and efficient for the organization. We applied speech to that. That's been a great business, continues to be a great business and one that we are going to really focus on expanding.
But what's happening now is also transformational, and it's the convergence of the things that you've seen here. Mobile devices are king. Something on the order of 60% of phone calls to contact centers are emanating from a mobile device, right? Not a big surprise. I don't even have a home phone anymore. I use my mobile device. But more and more people are using applications, right? Then, the Siri effect and in fact, the Nina effect has increased the expectations of customers that when they interact with these systems, it should be an intelligent interaction. It should be something that is natural, something that's human like. And that's emanating from the consumers and the customers. And when I meet with executives at the banks, insurance companies, they're coming to me saying, how do I deliver a virtual assistant experience to my users for my application, not just because it's cool, not just because they can do it but because the customers are demanding it. And the reason the customers are demanding it is because it's easier to use the application. It's easier to get things done. And that's really changing the game in this customer service world, and it's pulling into the other channels. If I get this on the mobile device, I should be able to get it through the phone call. I should be able to get it on my website. I should be able to get it on my ATM machine. And so they want to see this Natural Language experience, this virtual assistance experience wherever they interface with their customers.
A few years -- the Nuance on-demand system in kind of hosted services is benefiting from a trend that's also happening, which is people are much more comfortable with cloud services. 5 years ago, applications like salesforce, people were concerned about security and what about the latency? And what about my data? Is it secure? All of that's gone. We've proven it can be secure. We've proven there's no -- it's very fast. We've proven that we can actually deliver this in a way where the most important thing, the user experience is better because if we get the data and we can make the system more intelligent and it learns as it goes, it's a better way of delivering these services, and organizations are recognizing that.
And then finally, the voice of the customer is louder than it's ever been. If you deliver bad service, it's going to show up on product review posts or company review postings, Facebook, Twitter. People are going to know about it. And organizations know that it's far more expensive to recover your reputation and fix a problem than it is to deliver an elegant and nice and highly valued user experience upfront. So these things are coming together for something that's really, really good for the Nuance Enterprise business. This is set the stage the environment for us to be able to leverage what Nuance has as a company into the customer service realm in a way that's much more powerful than it's ever been. And that means a larger market opportunity, right? The Nuance Enterprise division isn't an IVR speech division. We are an automated customer service division. We are about delivering automated customer service experiences that people like and save our enterprise organizations money however they talk to the customer, however the customer talks to them. And we do that through the contact center, and we're now doing it through the mobile channel. We'll be doing it through the web channel.
All of this is synergistic. A lot of the on-demand result -- well, a good part of the on-demand result bookings at the end of last year came from our ability to show our customers, not only can we deliver this experience in the IVR, we can actually deliver it in a mobile device. And when people take a look at Nuance now, they're not looking at us as, "Oh, you're just one of many IVR people we could choose from." They're looking at us as an organization that is unique, unique in delivering this capability and all the touch points they have with their customers that are important to them. And we're able to prove that through relationships we are in. I mean to be able to be to come into a market and actually transform the way people think about automated customer service and to be able to kind of attack that opportunity, what better place than Nuance Enterprise.
We have such a strong position in the marketplace. We've got 3,000 contact centers, 8,000 speech deployments. We've got -- if you look at our on-demand service, approaching 80% of our customers have already been using NLU through the phone for the past couple of years. And those are growing.
And so to be able to come into this organization and to be able to move into new markets from such a position of strength with referenceable customers that have proof points and infrastructure and salespeople, now all the things that you need to be successful, I tell you I can't be anymore excited to be able to drive with my team this business into new areas and new opportunities.
And then add to that, what you've seen today. All of Nuance, right? So I go talk to an executive of a large bank and they say, Nina mobile is great. We want to be able to deploy that there. And can I get it in the phone system? And what about the web? And by the way, if a person can check their checking account or they can ask about mortgage rates so they can pay their bills on a smart application, what about the television set? What about the car? Can I bring that experience into the ATM machine? Right? And so you look at what Mike's team is doing, what Nuance is doing, what Vlad's teams are doing. We have the technology. We have the relationships. We know who Samsung is. We know who LG is. We know who Ford is. And we're oh, we're key partners with them. and so the ability for us to use those partnerships and extend that customer service experience into the car, into the television, as you look ahead, is very, very exciting prospect for Nuance as a company.
So Nina, this has just been amazing. We -- and so what is Nina? Nina is the virtual assistant for mobile customer service. It is an SDK that's iOS and Android and allows the enterprise organizations to take their existing mobile applications -- so not new work -- add virtual assistant capabilities to it. We provided them with a binary code that links to our cloud services for speech recognition, text-to-speech and NLU. We've given them source code for the persona, the Nina persona. And they can actually customize that persona. So Nina comes with 40 different expressions of sleep, awake, listening, those type of things. It's all designed for them so that they know that -- how that works inside the application. They could swap out, roughly speaking, the bitmap and have their own personality. They can use our custom text-to-speech voices to have their own branded voice around it, quickly integrate this into their application and deliver to their customers an experience that allows them to speak to the application and get what they want. And when you think about the mobile applications, they're pretty cool. You can click through and you can just say what you want, and you get what you want. "I'd like to pay my Visa bill in full next week from savings," and Nina understands that and does it for them. Or I might say, "Next week, I'd like to pay my Visa bill from checking," understands that. So our NLU understanding really allows us to have a very robust system that actually does what the customer wants it to do. And that technology has just created a level of demand for Nuance and engagement with Nuance that we've not seen in many, many years. It's provided us with conversations with organizations that, in the past, haven't talked us because they think -- they thought of us as an IVR-only type of solution. Now they see us as a broad provider of automated customer solutions, and it's providing more opportunities for us as we look ahead.
So as we look ahead for fiscal year '13, one key area is our Nuance On Demand hosted. We're going to be continuing to invest in differentiating that platform. As you probably followed along, we've added voice biometrics to On Demand recently. We've added Nina mobile to the platform. We're going to be expanding it with an even greater level of virtual assistant capability through Nina IVR in the future, as well as our Nina web solutions.
And in fact, early -- Vlad was speaking about a Chinese bank, a large Chinese bank that was using our system for SMS messaging if you remember that we got 90% automation out of that system, that's an early incarnation of what you'll see coming through our On Demand system. So a very, very great opportunity here, one that not only we can capture market share from the competition, but frankly, we can increase the number of organizations that want to move to the cloud because they see such the value that a cloud experience can bring to their business and to their customers.
Our Professional Services organization, I just -- this is such a competitive advantage to us. It's not trivial to design a really, really good customer experience. It's just not. And we know how to do it better than anybody else in the world, and we've got a large team supported by professionals. I think the tenure on our pro services division is 12 years on average, 12 years. These are people that not only love what they're doing, but they know what they're doing. And we're going to be strengthening this team. We're looking for growth opportunities in higher growth regions such as South America. We're going to be packaging up our offers even more to make the sales cycle shorter and really focusing on our cost of delivery but focusing, again, on the refined customer experiences we get because that's really what sells our product. It really is what makes the difference with our customers.
And of course, our on-premise, we've got very strong channel partnerships. I mentioned earlier Genesys, Avaya, Huawei and others. We've actually launched a new channel marketing program to put -- to get more leverage out of our channel, both on a global basis and a regional basis. We're also packaging up solutions there to help their sales team sell more quickly through into the market. There are some growth areas. We focused some selling around federal government and some -- and emerging markets. And we really do see a voice biometrics technologies in this segment starting to accelerate as we -- as that whole industry becomes more important.
And of course, Nina. This is the place where we have significant distance against our competition. And in many ways, Nina mobile in a mobile is the first-mover advantage that we have there. But our distance -- this isn't a product that came up in a few months, but this is something that the company has been investing in for quite a while, and it is a very, very sophisticated and robust implementation. We've had banking customers that have been able to implement prototypes, pretty extensive prototypes in a matter of weeks. We expect to see commercial deployments of this where you actually would be able to use it yourselves as users. U.S.A. is one of them at the start of the calendar year. And expanding that into the other channels is a very important advantage that we have as we move into '13.
So I'm very excited about what we can do with the virtual assistant space. It expands our market opportunity significantly, and it provides synergy back into and differentiation back into our IVR business that's beneficial to us. And it's already helping us. And our ability to deliver worldwide, I think, is something that's going to accelerate those successes as we go through the year.
So it's a very exciting time in Nuance Enterprise, and I thank you for your time today. And now I think we've got Kevin coming back up. Thank you, everybody.
Thanks, Robert. What I hope you've gathered through the first half of the day is that Nuance is developing and expanding set of technologies. These enable us to deliver new solutions that have the effect of expanding our addressable market and raising our competitive differentiation. Coming out of the break, we'll see how that applies to our health care business and is driving its growth for the next year. But in the meantime, let's take a 10-minute break. I have 11:15 Eastern Time. We will resume the webcast at 11:25 Eastern. Thanks.
All right. We are ready to resume. So if everyone can take their seats and let me introduce Janet Dillione, the General Manager and Executive Vice President of our Health Care Division. Janet?
Janet M. Dillione
Thank you. Thanks, Kevin. Good morning. Thrilled to have the opportunity to spend the next few minutes to talk to you about Nuance Healthcare. I'll spend sometime and discuss 2012 but I think more importantly, and with a lot of excitement, I'll also talk about the momentum and the opportunities we see going into 2013. Let's spend a moment here and talk about exactly who Nuance Healthcare is and what we do. Nuance Healthcare has established the position in the market based upon the technology foundation that you heard Paul speak about and you heard Vlad talk about and that's around speech and clinical language understanding solutions. We've established an indisputable market-leading franchise in front-end speech and capture. I'll give you some of those statistics in a moment. But that fundamentally puts us in a unique opportunity to take advantage of some things that are happening in the market. But again, we've made our reputation, and we're going to build on that reputation of delivering innovative technology that addresses complex problems in healthcare in very unique way. We did it with Dragon. We've done it with PowerScribe. We did it in transcription with eScription and you'll see we're fully and completely prepared to do it again in a different domain as we enter 2013.
We do it with customers and partners. Paul talked about it, you're going to hear about it from Steve. Our customer base is an absolute asset. As someone who came from a different part of the healthcare delivery system, I know others covet what we call our customer base. It's extraordinarily important to us, it's vital to us and we not only are proud of our customers but with customers like UPMC, we use them as a living lab. They are a living, breathing, very complicated healthcare delivery system. We work with them to deliver, develop some of these solutions that I'll talk to you about. And what I would share with you here is one thing. There's one metric, there's one gold standard in healthcare. It's the thing that everyone checks before they buy anything. And it's KLAS, and it's K-L-A-S, not C-L-A-S-S. And it's an objective measurement in healthcare that measures, did the vendor do what they said they would? Did they deliver what they promised? And again, the gold standard, every customer looks at it before they buy something. I'm pleased to say that across what we do with technology, we have 6 of our products in our portfolio are #1 in their KLAS category and one of our solutions is the #1 for 8 years in a row. Every year, we get #1, it continues to set the bar even higher.
So as we look at Nuance Healthcare, again, disruptive technologies addressing complexity, improving productivity, improving efficiency and while doing that, we've become one of the largest healthcare IT companies in the market of North America or globally.
Fiscal '12, very, very strong performance. Again, double-digit organic growth, a topline of $670 million, the three-year CAGR of 20%. Again, metrics, I think, we are proud of, I think, people and anyone in the market would be proud of. Along with the financials, we had a very busy fiscal '12. We had a lot of things that we needed to do. So I'm going to talk to you about those for a few minutes.
Let's focus on what we did in the areas of innovation because we are unique and that, again, we have very strong technology platform that we can draw from because of the things Vlad spoke about. We are very different than others that we compete with in the healthcare industry. In terms of innovation this year, we continue to take advantage of, again, our flagship position in capture technology. The fact that today, 450,000 clinicians will talk to Nuance. That's a fundamentally dominant market position to be in. And again, it delivers a lot of opportunity for us. Last year, I talked to you about, we are about capture, we are about capture anywhere. And this year, we've met that promise.
So we have dominance on the desktop. This year, we rolled out Dragon Medical on a mobile environment, onto a mobile platform. That's now being used by 350 development partners. And in the second half of fiscal '12, we introduced the cloud. So now we truly capture anywhere. Desktop, mobile device, on a cloud, and that cloud's important because there's a lot of parts of the healthcare system that are very far away, very loosely tethered to the delivery system. The famous rural access hospitals, clinical care providers who don't access the system that often, a very different use case and we think the cloud solution provides, again, tremendous opportunity. Most of our customers, we imagine, will deploy across those 3 different environments: The desktop, the mobile device and the cloud. We now have the solution that fully takes advantage of it.
Clinical Language Understanding remains a dominant, dominant place where we are investing in R&D. We continued to do that in 2012. We rolled out Clinical Language Understanding also to a development platform. We have 23 partners that are actually using our Clinical Language Understanding now to extract facts and to do things in their mobile applications and we will continue to do that and invest in that. But -- and 2 weeks ago, or maybe it was just last week, at RSNA, The Radiology Society of North America, that famous event attended by 65,000 people, where all things radiology happen, we introduced CLU to our flagship PowerScribe product line. And I'll give you, in a moment, I'll talk to you about how many folks use PowerScribe. Very, very, very popular, tremendous density at the booth. We had trouble accommodating it. But again, we've put that technology Vlad talked about, that fact extraction inside of the radiology workflow in a moment in time, it's very important. The U.S. federal government and governments around the world are focused on that area of radiology and imaging, why are so many high-end studies being done. And again, we've introduced technology to do that.
So again, these are the investments that we've made over the year. We successful deployed across the customer base and, again, are going to be the reasons why we've developed the customer franchise we have that I'll share with you in a moment.
I talked about partnerships. There is a dominant player in what we call the Healthcare ecosystem. And it's the EMRs. They are typically the largest enterprise investment a health system will make, they are the owner and the generator of most of the transactional activity inside of the health system. And it's very, very fundamentally important to the ecosystem in which we live and Nuance delivers complementary solutions.
We had 2 significant announcements this year. Due to that deep partnership that we've established, that collaboration that Paul talked about, you heard about it from Mike, you heard about it from Robert, in working with partners. Cerner and Epic both made tremendous commitments to us. Every mobile device in 2013 and for many, many years from now, every mobile device application from Cerner will ship with Nuance's speech technology. A tremendous testament to the scalability of our solution, to the reliability, to the dependability and the market presence. We made that press announcement approximately 2 months ago. And Epic, arguably the fastest growing force in healthcare IT, also has the press release and also made a commitment that their new natively-build mobile apps, Haiku and Canto, are rolling out with Nuance speech.
So you have Cerner, the largest healthcare IT company in the world; Epic, arguably one of the fastest-growing healthcare IT companies in the world, both making a commitment to Nuance. And I think, again, it's testament to the technology and a testament to how much we invest in partnerships and how seriously we take these is a very, very important part and a valuable part of our delivery channel.
And finally, in terms of 2012, in market expansion, we continued to invest in EMEA, actually we raised our investment levels in EMEA in 2012. We expanded sales coverage and we introduced 2 new products. We introduced our cloud-based speech. It's now available in 22 languages in the EMEA market. And we also, last week, introduced Dragon Medical 12, which has some of those speech improvements that Vlad showed you into the EMEA market. So we see the EMEA market beginning to migrate to more of a front-end speech use model, which we're prepared for. We also see, in the U.K. market, some changes happening with the NHS. A tremendous opportunity to bring our On Demand technologies into U.K. We began that investment in 2012. We will absolutely continue that investment going into 2013.
So as I look back at 2012. If I had to summarize, I'd share these statistics with you. Again, I think many, many folks that you will meet with in the market would love to stand here and share some of these statistics with you. The first one and, again, the one that makes all things possible is the fact that our Capture, our speech technology, today, 450,000 clinicians, a tremendously market-leading, completely differentiated position for us, one we take very seriously and we invest in heavily. 10,000 healthcare provider entities across the world have Nuance technology as part of their enterprise solution. We will this year process in excess of 5 billion On Demand transactions through our transcription engines.
And I spoke about radiology, and this is quite frankly an amazing statistic but today, in the U.S. market, because there's really only this type of the statistic for the U.S. market, 50% of every imaging study done in the U.S. today will be reported through a Nuance system. And that's just an incredible market position to have in radiology and, again, tremendous opportunity for us to grow and introduce new solutions like CLU into that customer base, but a phenomenal statement from the market about our PowerScribe franchise.
I mentioned we are #1 in KLAS for 8 year in a row and our eScription product line and the new metric for us this year, that I'll share with you, is 98% of the most wired hospitals in the U.S. are Nuance customers. So again, phenomenal statistics.
And as I look at this slide, as I look at these statistics, there's one consistent theme here, what this enabled us to do. And it was around Meaningful Use. So a few years ago, in the market, a lot of folks saw chaos and saw change and saw anxiety when Meaningful Use was introduced by the federal government. What Meaningful Use proved to be for us was an extraordinary tailwind. We had absolutely the right technology, the right solutions at the right time in a market with an unavoidable tailwind. And what this set of the statistics shows you is that the market views Nuance as a way to help achieve and attain Meaningful Use certification. We became that compatible, that complementary solution to help clinicians begin to use and adopt healthcare IT, and that was the fundamental bridge that had to be crossed with Meaningful Use. You didn't only have to own it. It had to be used. And that was a moment in time, again, where our technology proved to be different, unique, differentiated and really provided a phenomenal bridge for the customer base. So Meaningful Use, again, a little bit of chaos for the customer, tremendous tailwind for us. And we see further tailwinds for us and some chaos in the market.
So as we leave 2012, Meaningful Use, that which kept everyone awake for the last several years. If you talk to healthcare leaders today, there's some new things causing nightmares. The things that are keeping them awake. And there's really 2, if I had to simplify their lives for them. One is ICD-10 and the second is Accountable Care Organizations and the shifting mood around the payment models of the North American delivery system. And let me spend a moment on each of these.
ICD-10 is that change in the market where the business line which is healthcare, the way the centers of Medicare and Medicaid talk, the way insurance companies talk, the way payment systems talk, the way reporting systems talk, is called ICD-10. So it's not a heart attack. It not a cardiac infarction. It's not a coronary event. It's 410.2. That's not the way a physician speaks, it's not the way a nurse speaks, but it's the way the business of healthcare speaks. And that is changing. It's going from 50,000 codes to 159,000-plus codes. So it's numerically interesting. But what's really significant about it is every system in the healthcare IT ecosystem has a file in it somewhere called ICD-9. It just is everywhere. It's pervasive. And therefore, every system will have to be touched. This change, right now, is estimated to be, along with Meaningful Use, in the range of $100 billion of cost hitting the healthcare system and its payers and providers. So both are needing to go through this change. The ICD-10, absolutely keeping people awake at night and I'll share some more reasons with you in a moment.
October 1, 2014, be there, not changing, done, Kathleen Sebelius, federal government, Congress, boom. October 1, 2014, that's the date.
The second thing keeping folks awake at night is ACO, accountable care organizations. It's a buzzword. There's a capital A around ACOs, there's a lower case A around ACOs. Regardless of what had happened in the elections, regardless of ObamaCare, there is something happening in the market and it's around this idea that we want a more integrated, more homogenous healthcare system with less hand-off, less points of time, less transitions. And we also want a healthcare that gets paid for quality, for outcomes, for results, not simply for what was done. You've seen a lot of it in the press lately about corruption, fraud. The fact that, wait a minute, if you pay for what was done, isn't there a natural incentive for people simply to say they did more, or for people to do more regardless of whether it was needed. That theme is not going away. Whether ObamaCare happens, whether it happens as it's currently legislated, the payment system, the model, the delivery model, of the U.S. system will shift and it will be something around this idea of ACOs.
So those 2 things, ACOs and ICD-10, along with Meaningful Use, have 1 constant. One common theme at their core, and it's a beast called clinical documentation. And in the chaos of ICD-10 and in the chaos of ACOs and ICD-10, in that constant of clinical documentation, we see extraordinary opportunity. We see the same kinds of opportunities we saw when we addressed other complexities in healthcare. So let me spend a moment. We see a new market. We see a large market, a $20 billion market called clinical documentation. We see an opportunity to take advantage of what we've done with capture, that idea that we are with 450,000 clinicians today. That when people look at Cerner, when people look at Epic, when people look at McKesson and their other EMRs, they're going to be using Nuance to do that clinical documentation. We see an opportunity to take -- to leverage that capture position and more deeply extend into that much larger market opportunity called clinical documentation. Again, I'll give you a number of $20 billion. I'm going to walk you through in a minute what exactly that clinical documentation looks like.
But this idea of clinical documentation is extraordinarily complicated. There's a reason it's so messy. And we see an opportunity to once again be disruptive, just like we were with Dragon, just like we were with eScription. What Vlad showed you in terms of that Clinical Language Understanding technology, that ontology is fundamentally now more important than ever in what we want to do now as we enter this clinical documentation market. So let me walk you through and spend a few minutes. We've debated greatly about a graphic I would put up here, because it's almost impossible to graphically represent the chaos and complexity of this thing called clinical documentation in healthcare. It's both an output as well as an input as well as a process. What you need to know is this thing of clinical documentation as it exists in healthcare is represented everywhere that healthcare Is delivered, whether it be a physician office or whether it be a delivery system. The other thing you need to recognize is I went into any health system today, this would not be one linear graph with one platform or one technology. This would be a spreadsheet of vendors, of software, of technologies that someone would show you to say, this is how I do clinical documentation. That's the complexity, that's the fundamental chaos in the health care ecosystems. That's where we see extraordinary opportunities to introduce a disruptive end-to-end solution. Let's fundamentally again what we've done before and that's to bring productivity and efficiency through technology. Very complex, extraordinarily exciting. And I'm going to walk you through it.
First and foremost, what we can do, again, 450,000, we're at that point of documentation, we can go and address the physician. At that moment of clinical documentation, it's initial capture. We see an opportunity, along with CLU, to take advantage of a move that we did in 2012 with acquiring clinical knowledge to inject that clinical knowledge into CLU and deliver it at that moment of documentation. Sounds simple, but I want to give you a sense of this clinical knowledge that we now have as part of our engine. And as we speak in Burlington and other development centers, developers are migrating into our CLU engine. It took 25 years to build this core set of this clinical intelligence. It's now delivered across 500 customers but it's delivered as a service engagement, it's delivered as a training set. Because of our technology, again, that Vlad and Paul spoke to, we see an opportunity to take that knowledge and automate it and inject it in real time into the physician documentation process. No one else, regardless of any other analyst meetings you go to or anyone else who talk to you about clinical documentation, no one else can say it. Because that knowledge that is so unique and so disruptive and so darn exciting, quite frankly, for some of us who have been doing this for longer than I'd like to tell you. Just fundamentally, just tremendous opportunity. It's like as if the technology and the knowledge caught up with each other at the same time that the market was in a moment of chaos. A very unique moment for us. Thrilled and delighted we were able to take advantage of it. But again, at the point of care, a new intelligent system, I'm calling it that because I don't have a name for it yet. We are building it, but we will start fixing, we will start disrupting, we will start innovating clinical documentation with the physician at the front. Nobody else is going to be able to tell you that as you go through. Let's continue to walk through clinical documentation and other things that need to happen along this process in this value chain.
Next critical stop that's going to happen in this ecosystem called clinical documentation. At some point, I have to take that clinical language that I've accumulated. That chart, the fact that they had a myocardial infarction and a raised white blood cell count, oh wait a minute, did that mean sepsis? I better document sepsis. All that clinical staff going on, at some point, has to get translated. And there's a job actually in healthcare that just does that. There's a role, it's called a coder. And a coder is somebody who has actually quite a bit of science training, anatomy and physiology and they literally sit there and translate what the physicians and nurses have done into that business language of healthcare I talked to you about. So myocardial infarction becomes 410.2. There's a human that sits and does that. And that presents to you like, oh, isn't that pretty stupid, why can't you just take that phrase and translate it into a number. And that's because of some of the complexities that Vlad hinted at. There's an extraordinary variety of ways the clinicians describe things. And there's an extraordinary number of inputs that point to that simple thing of 410.2. And it's just now, again, that the technology is catching up to be able to solve this problem. So computer-assisted coding is the word. It is a tremendous market opportunity. It's a tremendous market opportunity because with ICD-10, that thing called a coder, the day of October 1, 2014, 50% productivity hit. 50%. Arguably, 2 years later, they say its grade will improve. The coders will get smarter. They'll still have the have 30% to 40% productivity hit.
So regardless of training, the other thing you need to keep in mind is, today, there are not enough living coders to address the problem of a 50% productivity hit across the health system. You can't hire them. Even if you could pay for them, which no healthcare system can. There's not enough alive. There's not enough in the school system. There's not enough getting trained. So we have a fundamental, again, opportunity. To us, when we look at that, extraordinary opportunity to some other's chaos. But we believe, again, we have the technology now because of what we've done with Clinical Language Understanding. We are able to extract those facts as Vlad showed. We can breed that knowledge that comes in because of our ontology. We can point to 410.2 and that tremendously exciting. But remember, we are also differentiated in our computer-assisted coding and you can hear others will talk about computer-assisted coding, no doubt about it. We have 2 things that differentiate. Number one, our underlying technology because we've been doing this for years and investing in it and this is a technical problem. This is not for the faint of heart. This is hard and it does require significant technology.
But second of all, we're already start better. We're giving them a clinically accurate document. We've addressed clinical accuracy upfront in that clinical accurate process I talked to you about, that intelligent system. So we're already handing the coder a better document from which to code. We're already handing computer-assisted coding a better document from which to derive codes. So again, same end-to-end solution, one integrated platform, one integrated set of technology and knowledge delivered. Now, I'm solving a second step in the process, I've addressed the physician, I've addressed the input. I'm now addressing computer-assisted coding, but I'm not done. I've got to finish up. I've now got to tell everyone what we did and how we did it. So I've got to generate reports, dozens and dozens and hundreds and hundreds, endless complexity in the delivery system, endless. They are called things like quality reporting. They are called things like performance reporting. They are called things like risks of mortality, severity of illness. Very complicated reporting that has to get generated. Again, same platform, same Clinical Language Understanding engine, same injection of a knowledge system allowing us to deliver that solution at the end of our process. So, we will have delivered the first end-to-end solution, one vendor, one platform, one technology, addressing the end-to-end issues of clinical documentation.
We will, again, fundamentally disrupt and do what no one else can do and we will absolutely deliver productivity. We will absolutely deliver performance improvements here. And most importantly, as this circuit begins in 2013, it's going to be just like front-end speech. It's going to be just like eScription. It's going to be just like our other solutions that depend on these types of technology. It's simply going to get better. Because the system will keep learning. The underlying technology will keep getting better, smarter, more precise and we will continue to be able to deliver productivity and efficiency improvements to our healthcare customers. So we are very, very excited. It's a new market. It's a $20 billion market. It's a tremendous growth opportunity for us. But again, when you hear me get excited, it's because I really feel we know we can bring tremendous value and do what we've done before in other solutions, which is be disruptive, be unique, provide improved productivity and provide efficiency in a way no one else can. A tremendously powerful position, very exciting as we enter 2013.
So across-the-board, we're excited, we have a fantastic portfolio. You're going to hear from Steve. We're going to continue to invest. We see dramatic opportunities for healthcare to continue to perform. We see dramatic opportunities for us to ever -- to become an ever larger enterprise partner for our customers. Again, we see an opportunity to be disruptive. We believe we know we've left 2012 fundamentally different than our historic competitors. We're simply not competing there anymore. We've entered a new, and a very different market in a new and very different way. We have the people. We have the technology. We've had this strategy, we've been working on this for over 18 months. This isn't a new idea. We are very excited and I'm delighted to be able to share it with you today. Thank you very much.
I'm sorry, I am supposed introduce, if I were to be accurate. Everybody, let me introduce Steve Chambers, our President and Head of Sales, globally.
Steven G. Chambers
Good morning. I'm going to try to talk to you about what you heard from the lines of business and how it gets expressed to our global selling organization and share with you little insight on our customer narrative. I know this gets very complex. We've talked a lot about the technologies. We thought it might be interesting to you to hear about how these technologies get expressed and what type of market demand we're addressing.
So as you heard from others, we left FY '11 into FY '12 with continued momentum. You saw the growth charts from Mike, Robert and Janet. Several factors contributed to that momentum. First, we applied more resources and expanded our coverage, both in the direct sales resource and also in the partnerships we established in areas of highest growth. We also saw tremendous demand for this concept of intelligent systems. More and more companies in this past fiscal year were soliciting our sales expertise, our sales reps and our corporate expertise to start to design, architect and look forward to see how those intelligent systems could further their businesses in future years. There were a profound number of dialogues we had around this concept of intelligent systems in FY '12 that bodes very well for future revenue production from the organization. So these deep collaborative customer partnerships are in evidence of those intelligent systems and the type of sales engagement necessary to talk about the intelligent systems in a way that specifically applies to Mike's global OEM customer, or Janet's IDN network of healthcare providers, or Robert's key enterprise banking and insurance customers.
A few words on the sales organization, we left FY '12 with 675 people in the selling organization. Our current plan going into FY '13, incorporating the recent acquisitions we just announced and also organic growth, adds another 100. So obviously doesn't contain what may be future acquisitions, contributions to the selling organization, but this is the current plan as we see it now, exiting FY '12, looking into FY '13.
To give you a sense of geographic presence, we have 55% of our selling organization focused in North and South America, primarily North America. We have 30% in Europe and the Middle East and we have 15% in Asia-Pacific and Japan. Across our lines of business in mobile, healthcare, enterprise, we had a very strong year-over-year revenue per head productivity increase at 15%. We'll talk a little bit about what the factors were that contributed to that increase, but it was an exciting metric for the selling organization as we engaged the market. What's really important to think about because I think, and particularly you just heard this from Janet, some of the ways our customers' feedback to us that Nuance offers unique value is in the form of those intelligent systems and the deep collaborative partnerships necessary to plan for those intelligent systems to be incorporated into our customers' products and services. I don't think we would be credible and I don't think that strategy would take root for Nuance and produce the gains it had if that stat on your right weren't in evidence. We have 7 years average tenure for our sales force in the domains in which they're selling. And we also have, based on scans we've done of the market, a low comparable voluntary attrition rate. People are doing well in Nuance sales. They're selling solutions the market wants. The tenure is growing and when the market demands an expert voice because of Samsung, because of Cerner, because the USAA is making multi-year bets, they want to talk to experts who can service their needs and bring them into the future. The obvious benefit from a revenue production standpoint, those deep collaborative relationships produce multiple years of revenue from those accounts and give us a very efficient sell-in mechanism to expand our footprint in those key accounts.
So some of the areas we did invest in FY '12 that are particularly important, one is health care sales. We grew our sales force, we focused in health care because there was such tremendous opportunity, and Janet explained some of the market dynamics that contributed to that health care opportunity. The geographic and language expansion has gotten a little, not quite the play I think it does when we're out selling these solutions in the market. When you go and have a conversation with an Intel, about a multiyear relationship of groundbreaking partnership that's going to redefine the interaction model between individual consumers and their laptops, one of the first things you'll encounter is how profoundly global their distribution is. We market voice technologies and natural language solutions that obviously rely on understanding the language spoken. So it's not a little attribute when we talk about languages supported, and in our range of solutions, that language support can go from 40 languages up to 90 languages in some technology categories. When these companies ask us how we can support creating a breakthrough new interface, a voice interface as you engage with your laptop, for example, the next thing they think about is, "Oh, and I want that SKU, not just to be in the U.S. and EMEA., I want that SKU to be worldwide, particularly in my growth markets, which often include Brazil, Russia and Africa." So the degree to which we're producing these languages, and Vlad's organization pushes that accuracy curve in new languages, profoundly impactful into our customer narrative in the field.
So we have a lot of selling models at Nuance. We have an e-commerce model, we have a variety of sales models with our various channels. We have direct sales partner-enabled sales. The dominant selling model for the company is solution selling, deep, consultative, solution selling. We invest to make our sales reps particularly efficient in having those dialogues and in our marketing to support those type of dialogue.
There are 3 companies -- I'd thought I'd give you an example, just from a sales customer relationship standpoint. There are 3 great examples of the deep collaborations, the deep partnerships and how we got them to come to Nuance and work on multiyear relationships which is that solution selling consultative dialogue. So let's start with Samsung, so you heard Mike talk about Samsung. Samsung is the largest consumer electronics manufacturer in the world and Samsung uses Nuance technologies phones, tablets, PCs, televisions and in the future, although I'm not saying exactly what they'll do, I'll just say in the future you will very likely see these technologies, our Nuance technologies infused into a variety of consumer appliances. Talking and listening appliances that are going to make their way into your homes.
There's a tremendous breadth to our relationship with Samsung, and we have over 100% year-over-year revenue growth, owing to that characteristic of a deep, collaborative, long-term relationship where we've enmeshed our technologies into their technologies and roadmap, where they've incorporated our software, but more than that, engaged us and our Professional Services and our user interface human factors expertise so we can design how their consumers are going to interact with content using the voice over a Galaxy phone or a Galaxy tab or a Samsung laptop or a smart television, or maybe in the future, your refrigerator. All that's possible. All that's going on with Samsung in this very important long-term partnership in Mike's mobile business.
In Janet's business, Cerner. Cerner was, by the way, not only one of the largest health care IT providers, as Janet said. Forbes awarded Cerner one of their top 10 companies in the United States for innovation this past year. This is a company pushing the envelope in health care, much like Nuance is. Cerner partners with us, we voice-enabled their EMRs so our proliferation through Cerner's very strong distributed selling organization. Our voice technologies for electronic medical record Cerner solutions is in evidence in the market. We had 50% year-over-year revenue growth with Cerner, and I believe that it's going to be accelerated in future years as well. And we also are looking with Cerner in the mobile domain. Janet mentioned we have a very special relationship where we -- I have a 100% attach or incorporation into Cerner's mobile solutions that are being delivered to market. In addition to the current Dragon Medical EMR integration, that's our front-end speech that we've sold in tight coupling with Cerner's EMR solution. And in addition to the mobile solutions of our Cloud-based speech used in conjunction with Cerner's mobile applications, we are going to advance our relationship with Cerner in key clinical language understanding areas, what Janet mentioned CDI, clinical documention improvement, and also we're looking to advance the relationship and embrace coding as well.
And in Robert's business, and this is a very significant company. USAA is an insurance provider to the military. The military sector of the U.S. government and the civilian population associated with the military as well. What's important about this partnership, and Samsung and Cerner has this characteristic too, is this company is meaningful in the customer service space. They just won J.D. Power's Company of the Year for Outstanding Customer Service. This company chose our solutions to power their self-service automation as they communicate to their customers. And owing to their commitment to customer service and their customers, they don't even call their customers, customers, they call their customers members. And they market to those members through a variety of increasingly Nuance solutions. Our revenue with USAA increased 100% year-over-year, and that as well is something that's going to continue in the future because they started with some technology run on their premise. They moved to adopt to what Robert described as our Nuance On Demand cloud-based delivery care system for all their telephony, customer care, member calls in, wants to check their balances, wants to check policy specifications, Nuance handles that call in our cloud. They are -- they need a customer so they recognize the importance of moving from premise to the cloud and out into the mobile channel. They also use us for outbound. So alerting their members about policy changes, about payments due. And they also use our product for a particular automated messaging capability called Missed Call Messenger. And there's more to come. Next year, I'm sure I'll be talking about 3 new things based on our dialogues now in this enterprise multiyear deep collaborative partnership between Nuance and USAA.
Now I talked before about a productivity gain increase we saw of about 15% year-over-year. This year, we're expecting approximately 10% and you might say, "Well is that decelerating?" And you have to understand this calculation assumes all the new hires we're bringing in to the company in sales, which is significant, and them going up a productivity ramp. So we are looking for strong double-digit productivity growth measured as revenue per head in the selling organization. It's going to come in large part from some serious investments we've made in sales force automation, in marketing automation. We've also redefined our sales process and methodologies, our forecasting systems and pipeline management, and we've also mobilized those tools, rendered them accessible to our distributed selling organization over mobile devices.
The combination of those improvements and investments in systems and process and methodology and access, we believe are going to deliver significant productivity gains next year for Nuance sales. My favorite slide because this is really all the people in the sales organization care about. They don't want to think about FY '12 anymore, we just brought over 900 people to Las Vegas for our worldwide kick off. And I have to say, just having worked in high-tech a while, I don't think I've seen an optimism and a spirit of, "Wow, I'm going to make a lot of money at Nuance this year." I don't think I've seen it in evidence, as much, ever as I did this past October, and it's in large part because of these. We just thought we'd tease out what we believe are the biggest growth drivers that, from a sales perspective, our folks are most excited about.
In Robert's business, it starts with Nina. Now, you heard what Nina was. I want to give you an example, just an anecdote of how Nina rolls out in the market in terms of our sales dialogues. We've been selling automated self-service for a while. Enterprise has a problem, Robert quoted 60% of consumers accessing their corporate partners, so me to a United Airlines or to an AmEx or to a Fidelity, 60% of the time on a mobile phone. In actuality, in some categories, like tourism and travel and banking, that mobile access rate is between 80% and 95%. So these businesses are aching for a better mobile solution, because they're visual interfaces, like a Bank of America's application, what was award-winning. Bank of America notes that it's just too complex. If you have to navigate through 15 menus to get what you want on a small screen, it's frustrating. There is this strong pent-up sense that the most proliferate consumer access medium that's personal, which is the individual phone, is being under-leveraged. This Nina product, and the way customers could access their information, just by voice and natural language drilling right down into the content they want from their business partners, is a profoundly important topic. It's getting us C-suite level dialogues in these accounts and a very significant, we believe, prospective bookings pipeline as a result. More immediately addressable is voice biometrics, let me give you another example of why this is important. You're go into an enterprise and you talk about their pressing problems. Invariably you hear, I feel there's abandonment in my mobile channel, can you give me a richer engagement model? We talk Nina. The next thing you hear is, "I don't know how to under -- I don't know how validate the identity of the people who are accessing their vital business information quickly in a way that's painless for them." They are frustrated because you and I forget our usernames. We forget our pass codes. We don't want to listen to security phrases, we don't want to be transferred to an authentication specialist agent and then get to our content. It takes too much time. So in a profoundly impatient consumer society, this is the answer. And you're going to see significant deployments of companies where you're going to call up, having preregistered your voiceprint and you're going to say at blank, blank, bank my voice is my password. And it will come back and say, "Hi, Steve, account validated. What would you like to do today?" And that speed of authentication makes a tremendous economic difference and user experience difference.
So those 2 areas, as well as more comprehensive engagements like USAA, that extend multiple years, that produce revenue over multiple time periods, are the focus and what's jazzing up our enterprise sales force.
In Mike's business in mobile, he spoke a lot about this but the customer narrative, going from embedded software to connected services is in every single one of our OEM deals. The headroom in the television market alone, given the penetration we have today, is huge. And if you listen closely to an interplay between Mike and Robert, you heard Robert characterize enterprises who say, "I've got to reach my customer through whatever medium they choose to access their banking information. It's going to be in the living room over the television." So there is this huge interest, solves the consumer problem with the television to navigate content faster, enriches the capabilities of a smart television and becomes a medium through which consumers are going to transact with their business partners.
It's a huge dialogue for us to have, and that connected service's richness is extending into our automotive narrative for model year '14, '15 and '16 cars, into our dialogues about televisions, into our dialogues about appliances like coffee makers. The degree to which intelligent Nuance systems power those devices is tremendously exciting to all these sales folks who sell to those particular device manufacturer markets.
Imaging has an interesting opportunity, which is a little bit unrelated to what we've discussed today, there is a key [ph] change going on in the MFP market where we're moving toward managed print services. That move towards managed print services, cloud-based authentication of users with MFPs and managed print services has put our document Imaging business in center stage in our dialogues with global MFPs. The Imaging sales force's purpose this year to perpetuate design wins to our current global partners of our managed print services and cloud-based document conversion services. And that will be a story to tell next year because that as well is another multi-year, deep, collaborative, customization services and technologies set of dialogues we're having with select customers.
I think if there were one rep population in October that was walking with a spring in their step, it was health care. And health care is a -- it's a particularly exciting place to be and the nexus of strong customer demand, solutions that Nuance offers, that in some cases don't have a strong competitive counterargument and regulatory, that makes it a great place to sell if you're a Nuance rep. Accordingly, we've increased our selling resources in these areas. We've increased our sales force in Clinical Language Understanding by 100%. So our Clinical Language Understanding specialists, we will have twice as many this year as last year and our diagnostics sales group, 70% more. With Dragon Medical, the front-end speech recognition, great product, 63% year-over-year growth, will increase over 30%. The sales professionals we have purposed to that front-end speech product enhance [ph] health care information management solutions or our server-based products.
So health care is particularly interesting. You add in the mix of strong market demand for intelligent systems. You add in new incremental markets for revenue production for Nuance of clinical documentation improvement and computer-assisted coding. You add to that ICD-10, that Janet mentioned and a variety of other federal motivators for compliance in the United States that are prompting consumption of the solutions Nuance provides. And finally, you overlay that with, we are the only company that provides that breadth of technologies. We've redefined the end-to-end solution and clinical documentation improvement. Our reps love telling those stories. Where better to work if you're a sales rep than a accompany that's mapping into a space that's active in its consumption and appreciates the differentiation of your story and what you're saying.
And just to conclude, we had a great 2012, looking to a great 2013. And we're in a unique position. So a common theme across all our lines of business that you've heard today, broad technical footprint. If you look competitively segment-to-segment, I think we stand unique in our end-to-end capabilities, just from a technology level. You add to that something very important, which is scale. We haven't talked about that here, but when Robert puts on billions of calls and Mike puts on billions of interactions in the cloud and glanet -- Janet puts on 450,000 clinicians, those companies have a very key buyer criteria, which is safety, proof, can you operate at the scale that I'm going to be throwing at your technology? For a Samsung, that's a big, big business. The idea that they're the #1 global manufacturer of android devices, that they need a partner who can support in their hosted solution, our hosted solution, their type of scale of consumer interactions, is profound. Same thing for lines of automated transcription. Same thing for calls handled in the contact center via automated self-service. So a key for success criteria for Nuance is this ability to support scale, languages, operational, networks, SLA adherence, common in our sales narrative. Then we have services and user experience differentiation, all are human factors experts. How often we dig deep into our corporate conversations in our customer conversations and talk about the user interactions? You saw Mike put up Wintermute and you heard Robo and that other pleasant voice. Those are just examples of personas. We get into those dialogues everyday. Every enterprise customer wants a unique experience for their system to communicate to their consumer. Nobody wants to sound the same. So the pro-services we offer, the human factor capabilities we have across cultures and all those languages, stands us a unique step. And if our sales people are here, they'd tell you I get in a room, we -- they talk, they, the customer talk about, "I want an enriched experience, I'm about consumer experience, I want my device to sound different. I want my bank to sound different, I want my television to behave and sound differently." That's when we really shine because we have the expertise to come in and give them a persona and give them advice on how culturally that persona might play in country market A versus country market B. And finally and narrowly into the sales area. I feel that this year we've really fine-tuned our sales machine. Our structural support for our market opportunities has been balanced to where the optimal opportunity is for revenue production.
In addition, introducing the systems and processes and methodology enhancements we've made puts our sales people in the right place to service this profound market demand and puts them in the right place with the right tools to efficiently deliver on our promise, collapse the sale cycle and create those year-over-year revenue gains with the world's largest companies who trust Nuance to power their most important devices and customer interactions.
So thank you very much. Hope that gave you a little bit of an idea of how we sell, and some of the messages associated with how we sell the lines of business. Now I want to introduce Tom Beaudoin, our CFO.
Thomas L. Beaudoin
Thank you, Steve. Welcome. And really thank you, all, for your interest in Nuance. Hopefully, you've seen today the tremendous opportunity that we have. And so I'll talk a bit around how that translates into our financials.
I'm really happy to be here today, especially after our fiscal '12. It included 27% EPS growth, 32% operating cash flow growth, and we believe that those record results were driven especially by our terrific activities in sales that Steve talked about, the technology underpinnings that Vlad talked about, all the strength in our business units. And as Steve talked about, what happens in many other areas within the company is a very strong focus on operational and cash discipline.
With that backdrop, where equally important is our focus on FY '12, in where we believe we have another significant opportunity. As you have seen today, through the presentations, we're excited about our market opportunity, the advances that we've made in our technology and the solutions that we are delivering. And we are confident that those are going to position us quite well to have a very strong fiscal year in 2013. In fiscal '12, Nuance delivered 25% total revenue growth and 12% organic growth. We guided that fiscal '13 organic revenue growth would be similar to our fiscal '12 organic growth. As today's presentations have illustrated, we provide our solutions through many, many, many different delivery models and sales models. This has facilitated our ability to both grow the license revenue line, which is the blue line here, and our On Demand revenue represented by the green. We will have, as you've heard today, we will always have multiple revenue models with license revenues continuing in all of our segments. But specifically related to On Demand, revenue has grown at a 67% compounded annual growth rate from fiscal '06 to fiscal '12. We talked about Mike having really no On Demand revenue only a few years back. From fiscal '11 to fiscal '12, the growth in our On Demand was 30%, for a total revenue at last year in our On Demand solutions of about $496 million. In fiscal '13, we expect continued growth from On Demand, driven by the usage models with our mobile contracts, especially with those customers, Samsung and Intel, renewed growth in our enterprise On Demand, revenue and bookings and continued success in our health care On Demand offerings. And I'd just like to remind everybody that with those On Demand solutions-type revenue, those revenues come in over time but are strengthened by those deep relationships that Steve talked about with those particular partners, and they have a tendency to build upon themselves as we continue to add solutions, technologies and capabilities around those early wins and those early successes. Our business model, which allows us to grow both the license revenue, the embedded product and the licensing products, combined with that strong On Demand solutions revenue is relatively unique in the software industry.
Over the past 6 years, On Demand revenue has grown from about 6% of our revenue to 29% at the end of fiscal '12. I would like to point out that again, license revenue, as represented by the blue line, has continued to be our largest revenue stream. So despite a very, very strong growth rate in On Demand solutions and hosted activities, our license revenue and the solutions that people are requiring from that has continued to be very strong. And our ability to simultaneously grow the license and On Demand revenue is really, I think, a testament to what Vlad talked about in his technology presentation and then the solutions that you saw through the business unit presentations. And as Steve talked about through the way we go to market and the way we offer those varying solutions to our particular customers.
In fiscal '13, we expect a continued mix shift to those on demand, as you see those exciting solutions that we've talked about that will be rolled out in healthcare, renewed bookings growth in enterprise On Demand, including Nina, and the new On Demand streams that Mike talked about in mobile.
I think that it's important to point out that despite the mix shift over the last few years to On Demand, since 2006, our total gross margin, which is reflected in the black line has held fairly steady. In addition, our gross margins for PS and On Demand, represented by the green line, has improved by 12 percentage points since 2006. PS and On Demand gross margin have benefited from a greater proportion of the revenue coming from those On Demand solutions. We used a lot of those deep services solutions in our On Demand and so we still have a Professional Services revenue line, but we also have the deep solutions customization that ends up in our On Demand. But as we get more efficient, we get more productive, you heard Robert talk about repeatable solutions, we've been able to grow those gross margins.
As we guided in our earnings call, the revenue mix in FY '13 is expected to decrease total gross margin by about 100 to 200 basis points.
Our management team, as you heard today, continues to focus on the operational programs and the go-to-market programs to increase the value in price of those On Demand solutions to improve the productivity and the efficiency as we deliver those, and to contain the costs of delivering those activities. All those activities should provide for margin improvements in the future.
With respect to operating margins, we have a demonstrated track record of driving efficiency, productivity and leverage in our infrastructure and support groups. Both within the company, and also through our acquisition synergies. This focus, along with strong license revenue, has resulted in operating margin increases of 3 to 4 percentage points from fiscal '06 to fiscal '09, and 1 to 2 percentage points per year since then. As you saw in today's presentations, we have a number of strategic opportunities to expand our markets, and our plans include accelerating the investments in R&D. We expect these investments will fuel organic revenue growth. In addition, while we continue to invest in sales and demand generation, we are realizing sales efficiency, especially as we expand our On Demand revenue. During FY -- during 2013, the impact of the mix shift on gross margin, the strategic investments we discussed throughout the day, and the continued efficiency and productivity activities that we're all driving, will provide some modest pressure on operating margins for fiscal '13. However, the medium and long-term return on those investments, plus our continued productivity initiatives should provide operating margin opportunity in the future.
I think this is a pretty impressive chart. Our non-GAAP net income grew at 42% compounded annual growth rate from fiscal '06 to fiscal '12. During that period, EPS grew at a 29% compounded annual growth rate, and I'd also like to point out that the EPS growth over the last 2 years has benefited from a significant reduction in our share count. So just quickly, you can see through the chart in FY '11, our share count growth was about 4.4% and in fiscal '12, it was 1.5%.
A critical element, as we talked about in last year's Analyst Day, was our efforts around a global tax strategy, and part of that strategy included the relocation of our international headquarters at the beginning of fiscal '12. This change in how we conduct the International business, coupled with continuing tax planning, enabled us to achieve a low tax rate in fiscal '12. Accumulated NOLs, stock-based compensation deductions and continuing operational and tax planning efforts will allow us to maintain a low cash tax rate. We expect the fiscal '13 tax rate to be high-single digits and for the rate to grow gradually over the next few years, to approximately 20%.
I'm pleased with our operating cash flow performance. Cash flow from operations grew at 40% compounded annual growth rate from '06 to '07 and as noted earlier, grew 32% in fiscal '12. Over the last 3 years, cash flow from operations as a percent of non-GAAP net income has consistently been between 80% and 85% on an annual basis. As we've described, the difference is driven by acquisition and restructuring-related expenses, normal growth in working capital requirements as the business continues to grow, and tax deductions from stock-based compensation, which I would like to remind you are reclassified to cash flow from financing, and therefore has 0 impact on our total cash balance.
We expect the ratio to hold on an annual basis with some quarter-to-quarter fluctuations, primarily related to the timing of customer receipts and payments.
Nuance has dramatically deleveraged over the last 5 years through growth in revenue, EBITDA, and cash flow. We exited fiscal '12 with a cash balance of $1.1 billion.
Our August and October bond offerings were extremely successful in the market. The deals were significantly oversubscribed and were priced at a very favorable interest rate. These bond offerings did increase our leverage somewhat but leave our leverage ratios still well below fiscal '07 and '08. And despite the bond offerings, we were able to maintain our debt rating while raising this additional capital. We feel the company benefits from our current debt rating, which enables us access to capital at very reasonable rates.
I'd like to just point out that through these financing activities over the last couple of years, we have created a very attractive maturity schedule over the next 8 years.
So in summary, Nuance has delivered very, very consistent revenue growth, both organic and as reported, with an increasing proportion of our revenue coming from On Demand and other recurring revenue streams. As you heard through the presentations today, we have an opportunity for sustained double-digit organic revenue growth. And a continuing deep relationship with those customers that provide those stronger and growing On Demand revenues.
Our solutions serve large and growing markets. Nuance and our customers are creating new use cases for our technology, especially as our value proposition expands from speech recognition, to Natural Language Understanding to intelligent systems. These expansions and our solutions in use cases have increased our addressable market and raises our competitive differentiation. We believe it is the right time to increase our investments in technology, sales, demand generation and services to improve our ability to serve these expanded markets.
We believe these investments can contribute to improved growth in the second half of fiscal '13 and beyond and we continue to maintain a strong set of discipline and focus on improving productivity and efficiency, optimizing our margin and increasing our cash flows. We have a balanced capital structure affording us the ability to finance our growth at a reasonable cost to capital. We enter the new fiscal year with a strong technology-to-market position and very strong operational capability that positions us well to achieve our financial objectives.
This is a really fun and exciting place to work and, as you heard today, we have a great management team with an extraordinary set of technologies with some of the best customers in the world.
So thank you very much for your attention and we'll move to Q&A.
Okay, we're ready to begin.
John F. Bright - Avondale Partners, LLC, Research Division
John Bright, Avondale Partners. Paul, we've heard an impressive presentation today about intelligent systems. Talk to us about what you need strategically to fill holes from an acquisition standpoint looking forward?
Paul A. Ricci
Well, I don't think we have specific needs with respect to acquisitions to execute on this strategy. There may be opportunities for acquisitions to complement the strategy. There may be opportunities to broaden our solution set in one of our specific markets, to accelerate our position in a specific geographical market but I don't think there are missing elements that we're reliant upon in order to execute on this.
John F. Bright - Avondale Partners, LLC, Research Division
What should investors think about in the upcoming year? I think you mentioned on the conference call, limited acquisitions. And then in the presentation, I think you mentioned something about $50 million in revenue to add in the upcoming year when you were talking about healthcare. Is that -- did I have that right?
Paul A. Ricci
Yes. Well, I was referring specifically in healthcare to a couple of small acquisitions that are in flight and I expect those will be closed. And I think that it's difficult to know. As you know, we get asked this question every year and every quarter, it's difficult to know because some acquisitions present themselves opportunistically. I think though that the likely picture that you should have is one in which acquisitions are playing a relatively lesser role in the overall composition of our growth as we look out over the next year. And the typical kind of acquisition, I think, you should expect, will be on the smaller side. The kinds of acquisitions I'm referring to may be acquisitions of $50 million or $100 million or $200 million, relatively smaller acquisitions, and I think with somewhat less frequency.
John F. Bright - Avondale Partners, LLC, Research Division
One question for each of the heads of the respective segments. Again, intelligent systems seem very impressive throughout the day. Can you talk about your competitive lead in this theme as you look forward?
Paul A. Ricci
So on my side of the Mobile business and I -- lead, it sounds very binary, but I think we have very strong lead. The depth and breadth of what we're doing, the focus on voice recognition specifically, and the breadth of our use cases in languages times the different places where we apply this, in cars, on phones, tablets, living rooms, it's quite a broad and diverse portfolio, and I think we got a real lead.
And -- Robert, with the Enterprise division, I'd agree with that. And again, I think lead is maybe too limiting. We've got significant market differentiation advantage in what we're doing. And it's a combination of the intelligent systems backed by the language breadth and others. In my particular area, the Enterprise division, we really see Nina as a first mover. If you look at what's been out there, there's nothing like Nina that came before and, in fact, have not seen anything really since. And the combination of that, the synergy we have with that across the multiple channels is another advantage that we have. And then finally, the fact that we own the full stack. If you look at a different place, we have different competitors, no real single competitor in the different segments we have. And we have control of our speech recognition, our text-to-speech, our [indiscernible]. We can really get into this and have those close relationships with our customers and deliver exactly what they want, where others really have to license and cobble together their solutions. And so I think it's the combination of those things that give us a leadership position, if you want to phrase it that way, that we feel very confident about.
Janet M. Dillione
I'll just add, I mean, it's significant. We are differentiated. If I look at the multiple vectors from a technology perspective, rules-based, statistical learning that the machinery itself we know through metrics, through competition through the market, we're just -- we're different, we're better. We have an advantage right now. And we've made a significant move in 2012 to differentiate now in content. We already have ontologies that Vlad talked about, and healthcare ontologies are desperately important, desperately important because of the complex language. And we've added to that, again, with the knowledge base that we've added this year. And third, I would add is the power of the ecosystem of healthcare. Those partners, those EMRs, are desperately important and we have worked with them and they've helped us figure out a way to be very complementary, to integrate very well and to make it seamless for the customers. So I would be -- I am very happy I'm at Nuance and not trying to compete with me.
Scott Zeller - Needham & Company, LLC, Research Division
Scott Zeller from Needham, healthcare question. Regarding ICD-10, could you give us a sense of how far along we are into the buying patterns given the 2014 deadline? If you can just broadly frame out the urgency around spend. And when we might hit the high point of the urgency around spend regarding...
Janet M. Dillione
I think a good way to think about it is that we are in an early wave of buying a CAC, computer assisted coding, which I think everyone thought was the single solution to ICD-10. Went through it and is in a hype cycle. I think the early market is now full of CAC. CAC is important but insufficient to solve the larger problems. As the market actually pulled back, #1 the government delayed it from October 1, 2013 to October 1, 2014, so the buying cycle reiterated itself and pushed back a year, if you will. I think you should think of 2013 as the budget year, right? Everybody in healthcare will have a budget line item called ICD-10. So that's why we made the moves we did in this past year, that's why we're pushing so hard. We have solutions in the market and you're going to see us every month in 2013 continue to push them out there because the buying cycle is through 2013. Most customers would like to be in a dual ICD-9, ICD-10 environment somewhere Q1, Q2 2014 as they get ready for October.
Shyam Patil - Raymond James & Associates, Inc., Research Division
Shyam Patil from Raymond James. Paul, to follow up on healthcare, $1 billion in revenue assumes, I think, about $330 million incremental. Exactly how much of that's organic versus inorganic? And do you need the $50 million acquired revenue to hit that target?
Paul A. Ricci
Well, the arithmetic is that we are projecting about the same organic growth rate we had this year on our core business. We're layering on top of that the revenues from the 2 acquired assets, Quantim and J.A. Thomas. And we're assuming that there will be something in the range of another $50 million in contribution perhaps, perhaps a little less, perhaps a little more. And that gets us to a number that's somewhat over $1 billion. I think that's the basic arithmetic.
Shyam Patil - Raymond James & Associates, Inc., Research Division
Okay and then this is for Tom. I know it's early in the fiscal year, but when you look at the guidance for fiscal '13, what are some areas on the top line where you think you've been conservative and perhaps have a good chance of outperformance?
Thomas L. Beaudoin
Well, I think as you saw today, there are a number of areas of opportunity. And depending on how quickly some of those things move -- but some of it's offset a little bit by the ratable nature of some of the new solutions. There's a lot of the solutions that you heard today are delivered on demand through our hosting solution. And some of them are mixed, and some of them take some license in. So I think as we talked about at the end, and I said at the end of my thing that if some of these activities that we talked about today can overperform a little bit, then we might see some upside in the second half. But we'll just have to wait and see.
But we're not using today as an opportunity to raise our guidance, let's be clear about that. We gave you a revenue range recently. It's an ambitious revenue target and we'd encourage you to continue to think about that range as we gave it to you.
Brent Thill - UBS Investment Bank, Research Division
It's Brent Thill with UBS. Steve, I look forward to that talking refrigerator in my house at some point. I think you're going to expand your sales capacity by about 15% into the coming year, can you give us a sense of where you're already pushing that capacity, is it split across the divisions, or is there one particular division that's going to get that resource?
Steven G. Chambers
Significant in healthcare, in clinical language understanding in particular, and the front end speech of Dragon Medical. And I'd say that, that is the area of emphasis that we see. We're in a very good position for coverage in mobile and Mike's business in OEM, very good position in Robert's. Healthcare, we're going to really increase our coverage through partnerships and expansion of the direct sales force.
Brent Thill - UBS Investment Bank, Research Division
And a quick follow-up for Janet, just back to the healthcare question, ICD-10. Can you just give us a sense of what these early cycles look like? I think we're all looking for, is there a penetration rate right now? Or you hit a certain percent in ICD-10? Or is it, you talk in the first couple of innings in terms of the penetration. What these sales cycles look like, and are you dragging a lot more infrastructure along, a lot more cross-sales, along with the sale in your business?
Janet M. Dillione
I would say that, to the latter point of your question, I mean the early experience that we're having is these are very broad sales. They are that end-to-end system that we talked about. I can't think of a single instance where we're in just talking about computer-assisted coding. So I would say that's the first. And it's just simply very early. And as a matter of fact, at HIM [ph], in October, which was right after we announced Q and J, as we seamlessly called it inside, it was amazing how many customers had stopped. They had purchased something, had unrequited love, it wasn't quite what they expected and were going to go back out to market because the federal government had delayed. So we are very early in the buying curve. I mean we are literally -- I myself was out to 3 last week. You're talking about -- we just done our fiscal '12 budgeting, we see it in 2013. They're going out to our space. So it's early in that buying cycle. But again, not -- people -- I haven't seen one yet where it's just about computer-assisted coding.
Brent Thill - UBS Investment Bank, Research Division
Can I just follow up real quick? Just in Q4, there were a little question marks around your organic growth, perhaps because you pushed the date out or the date was pushed. Did that cause some delays in terms of what you saw in your business unit, where details are going to be somewhat bulky. It's hard to time the exact time of the close, but...
Janet M. Dillione
I mean we finished with what we've forecasted as double-digit, 11% organic growth. I think, as Tom said, we began to see some sort of -- all of our new solutions are cloud. So we've begun to see a little bit more ratable revenue in the healthcare line than we'd seen in previous years. But we didn't see -- there was nothing unexpected in Q4. No. We landed where we expected to land. We landed on the top line that we expected to hit.
Shaul Eyal - Oppenheimer & Co. Inc., Research Division
Shaul Eyal with Oppenheimer. A question for Tom. Tom, internally, what's the exact definition for organic revenue? Is it kind of x acquisitions that were bought over the past 12 months, 18 months?
Thomas L. Beaudoin
We've been very consistent for a long time about our definition of organic growth and it's published in all of our results and it assumes that we had the acquisition for the previous period of time that it's being compared against.
Shaul Eyal - Oppenheimer & Co. Inc., Research Division
Question for Mike. I think last year, when you provided your presentation, I think you talked about that time being embedded, I think in about 40 different Android models at that time. What's the current number, in case you have it?
I don't have a specific number but it's huge. It's huge. We should also mention, of course, that what has happened in the time since last year is that the market share is consolidated around a much smaller number of Android devices. So the number is less relevant than the specific devices in which we're present.
Rich Close [ph] from Columbia Management. Janet, can you talk a little bit more about the 2 recent healthcare acquisitions, how the integration has gone, what you guys are thinking in terms of timelines for when you're going to have an integrated product out there available for your sales force so that we have something to kind of look forward to watch for? And then just in terms of your timeline for that product versus what the health care system is kind of looking for their timeline. You talked about 2013 being the budget year. So what are they kind of looking at in terms of their timeline for when they need a product from you?
Janet M. Dillione
Sure, in terms of the Quantim and JATA platforms, with Quantim, we got a very broad, web-enabled, built from the ground up, newest clinical documentation HIM platform in the market. So that's exciting. It's very open technology, very complementary to the portfolio. We have -- we are building computer-assisted coding on top of that platform. We will go to market. We have a product now. We have an alpha release, we have an outpatient ICD-9 product, we have been in several customers. We have a ICD-9 and expanded code coverage release on outpatient going to our customers in February. I mean, ICD-9 inpatient, ICD-10 inpatient, ICD-10 outpatient following. So every quarter, you'll expect to see further customer deliverables from us. But again, right on top of -- it's completely integrated. On top of that, that coding platform that we have today. Already, today, our sales force is selling full solutions, alright? The RPs [ph] are coming out. They're not looking for a point in time. They want an end-to-end solution. It's one of the reasons why some are stumbling in the market, because the customers are saying, wait a minute, my computer-assisted coding has to be integrated with my clinical document improvement process or I run the risk of fixing things twice. I run the risk of solving the same problem. So that's very good news for us. On top of that, so I'm going to have my computer-assisted coding, online-integrated end-to-end platform. Next to that, I have my content from JATA that will support this clinical documentation specialist. And in Q1 of 2013, that intelligent device I spoke about where we had that clinical knowledge presented at the point of documentation where we're driving clinical accuracy, that will be integrated on the platform and, again, we'll be in beta in the February timeframe. So we find it to be right now. Again, there was a wave of buying when the date was October 2013. Again, as I said at HIM, what we saw -- and Paul was there as well. I mean, many customers made an initial decision and are pulling back. It wasn't quite what they wanted or quite what they expected. So some of those are back in market but it's very busy right now. Once we expanded the sales force, it's very busy. It's going to be busy through fiscal '13. But again right now, you got to think, people -- most of healthcare providers are coming out of end of '12 as their close of their fiscal year. So now they're going into fiscal '13, which is their fiscal '13 operating budget. They've got a line item called ICD-10 and that's the capital, that's the budget they're going to go to market for and that's the one we're going to go compete to get.
Nandan Amladi - Deutsche Bank AG, Research Division
Nandan Amladi, Deutsche Bank. Question again on healthcare for Janet. With these 2 new acquisitions you just made, does that change your revenue model at all in terms of license or on demand? There seems to be maybe potentially higher services content with JATA and Quantim.
Janet M. Dillione
I think that would be the only difference, is that with JATA, with that clinical knowledge, we absolutely will see incremental services revenue for us. Quantim is pretty consistent with what we see in other diag [ph] and our capture product lines in terms of revenue -- types of revenue in terms of the license and services. But with JATA, it's not only incremental growth in services, it's a different type of service. It's a very, very deeply and uniquely differentiated clinical service. This service is only delivered by doctors and nurses. So we are excited, it's a very high-end service. And we believe it's the beginning of -- we see dramatic growth in services. Steve talks about incrementing on investments and sales. We, as a division, we have a significant investment in our services in terms of people and products that we are putting into the market in 2013.
Nandan Amladi - Deutsche Bank AG, Research Division
And a short follow up, if I might. This might be for Tom. We've used the transcription line volumes as a key metric in healthcare. As you move up the chain and offer a broader solution, does that still remain an important metric? Probably it remains relevant but...
Thomas L. Beaudoin
Yes, it remains relevant but a lot of these other solutions -- and we'll continue to look at the metrics that we publish to try to help you to understand that going forward.
Daniel H. Ives - FBR Capital Markets & Co., Research Division
Dan Ives, FBR over here. First off for Mike, could you talk about -- I mean it's good information you gave. Will you just talk about how it translates to revenue? Do you feel like the pricing power is stronger for Nuance today into some of these mobile carriers? Or is maybe a year or 2 ago the financial terms were not beneficial to Nuance? Where as the newer deals we could view as more revenue opportunity that we've seen in the past?
Well, for several years in a row we've increased the sophistication of our solutions, which has been able to drive higher prices on a per device basis and we continue to do that. The nature of the solution is more sophisticated and more complex and we continue to drive that up. These NLU systems, connected cars, these are totally different pricing schemes with long-term relationships and so it's in flight.
Daniel H. Ives - FBR Capital Markets & Co., Research Division
But does that include a more variable component versus the fixed?
We have very complex contracts, many variations of contracts, and it's hard to just say a simple answer to that question. Some are fixed, some are variables, some are a hybrid.
Daniel H. Ives - FBR Capital Markets & Co., Research Division
Okay, for Janet, just -- you talked about the healthcare opportunity and obviously, you guys have proven it out over the last few years. But just given how aggressive you were on acquisitions this year, especially on healthcare, just walk us through the methodology why you need to continue to acquire there given the total adjustment market opportunity is pretty large, right? So just maybe just walk us through, why acquire rather than just do it yourself?
Janet M. Dillione
I think, as Paul said, I mean my operating plan is to execute on what was done in 2012 and to deliver what we committed to the board and to the shareholders of what we did in 2012. And we will be opportunistic in 2013, we're always going to be. If there's a way to augment technology, if there's a way to expand geographically. Some markets are less advanced than the U.S. right? So there's some growth opportunities in some different areas in the globe. So we'll continue to be opportunistic. But the bulk of my time right now is spent operationalizing and executing on what we did in 2012.
Paul A. Ricci
Just to reiterate on that, I wanted to emphasize a point. Our healthcare plan is an organic plan. There have been a couple of small opportunities that have arisen, primarily around geographic expansion, some customer segment expansion. Those typically make sense for us and we do them if, economically, they make sense, and we don't if they do not. And I think that's what you should expect for this year.
This is for Vlad, I guess. A large percentage of your employee base are transcriptionists. Does the technology get to the point where that number of transcriptionists as a percentage of revenue, however you want to look at it, declines and eventually, will become much more automated?
Well, so I can answer I think half of that question and the rest is for Janet. So my job is indeed is to try to minimize the work left for the correctionists who will become essentially proofreaders for the automatically generated drafts. And the performance or productivity in yield on that task continues to go up. So Janet, do you want to talk about the labor side of it?
Paul A. Ricci
I might add a few comments before Janet jumps in. So I think you have to look at the sweep of this over a period of the last half decade. We fundamentally disrupted the manual transcription industry with our technology. And today, we displace about 90% of the labor content in doing the task. So it is profoundly automated today. Because of the high level of quality threshold, there's some editing and verification that goes on and there's still plenty of unautomated lines out there to continue for us, to continue to sell to and attack. But the transcriptionists we have represent a very small labor contribution across the 5 billion lines annually that we're processing. And the majority of lines today, still are not being fully -- not utilizing speech, and so there remains a lot of opportunity. So the question is not so much whether the number of transcriptionists goes down but what is the proportion of automation that we're doing in the lines that we're actually processing. And we will get more and more leverage out of that labor content across more and more lines over time. Janet, do you have anything to add to that?
Janet M. Dillione
No, I think that's great.
Jeffrey Van Rhee - Craig-Hallum Capital Group LLC, Research Division
Jeff Van Rhee, Craig-Hallum. Maybe Paul or Tom, you talked about the gross margin declined 100 basis points to 200 basis points. You've got a mix shift going on. Could you get a little more specific and quantify what kind of shift you're looking for? Give us a little finer point there? Tom, you gave us some good slides there on the mix shift, but just to expand on what exactly the magnitude, the shift you're looking for here?
Thomas L. Beaudoin
As we've talked about, we're building out very complex solutions that's across many of these groups and a lot of those investments require some level of services and R&D and capability and then as the revenue flows, it depends upon how that revenue is taken and it depends how the costs flow against that, right? And so as we see this shift to more cloud-based services, which drives more On Demand revenue, which therefore drives more time-based revenue, we see a little dip this year, because of the nature and the size of those. But as I said, and as we've talked about, there's a lot of investments going in and driving efficiency and productivity. Paul just talked about it as we drive up the efficiency in the healthcare transcription. We need less labor, we get more technology across that labor base.
Jeffrey Van Rhee - Craig-Hallum Capital Group LLC, Research Division
Just to follow up, I mean, is it possible that you could quantify that? I mean, you gave the percent of revenues each year is from the On Demand channels. So what are we looking at for '13? How rapidly does that percent change as a percent of mix?
Paul A. Ricci
I don't think we can give you more detail than that.
We're done? Okay then, we thank you, all, for joining us today and we look forward to continue the discussion at lunch. Thank you very much.
Copyright policy: All transcripts on this site are the copyright of Seeking Alpha. However, we view them as an important resource for bloggers and journalists, and are excited to contribute to the democratization of financial information on the Internet. (Until now investors have had to pay thousands of dollars in subscription fees for transcripts.) So our reproduction policy is as follows: You may quote up to 400 words of any transcript on the condition that you attribute the transcript to Seeking Alpha and either link to the original transcript or to www.SeekingAlpha.com. All other use is prohibited.
THE INFORMATION CONTAINED HERE IS A TEXTUAL REPRESENTATION OF THE APPLICABLE COMPANY'S CONFERENCE CALL, CONFERENCE PRESENTATION OR OTHER AUDIO PRESENTATION, AND WHILE EFFORTS ARE MADE TO PROVIDE AN ACCURATE TRANSCRIPTION, THERE MAY BE MATERIAL ERRORS, OMISSIONS, OR INACCURACIES IN THE REPORTING OF THE SUBSTANCE OF THE AUDIO PRESENTATIONS. IN NO WAY DOES SEEKING ALPHA ASSUME ANY RESPONSIBILITY FOR ANY INVESTMENT OR OTHER DECISIONS MADE BASED UPON THE INFORMATION PROVIDED ON THIS WEB SITE OR IN ANY TRANSCRIPT. USERS ARE ADVISED TO REVIEW THE APPLICABLE COMPANY'S AUDIO PRESENTATION ITSELF AND THE APPLICABLE COMPANY'S SEC FILINGS BEFORE MAKING ANY INVESTMENT OR OTHER DECISIONS.
If you have any additional questions about our online transcripts, please contact us at: firstname.lastname@example.org. Thank you!