Less than four months after I attended Microsoft's (NASDAQ:MSFT) Data Insights Summit in Seattle and proceeded to gush, post-Data Insights Summit, about the potential predictive and dynamic learning powers of Microsoft's Power BI - a subject that I interviewed Peter Van Der Made, CTO and interim CEO of California-based Neuromorphic-IP licensing company Brainchip Holdings (OTC:AZKLF)(BRN.AX), about (see below) - it isn't Microsoft that's out with a major announcement regarding either predictive or dynamic learning but Brainchip. At this point, given what I'll breakout in the note below, I think Microsoft should consider a JV or entire company takeout of Brainchip or another provider of neuromorphic capacity. I just think that Microsoft would be well served, again given the dynamic broken out below, investing in "purchasing R&D" for immediate deployment into its BI platform - specifically that of neuromorphic, AI, and/or machine learning. At the very least, Microsoft should figure a way for integration of neuromorphic technology into the Power BI platform. Microsoft peers Dell, SAP, and Oracle have already deployed M&A dollars to these areas (for their respective BI platforms) and I believe it would serve Microsoft investors well for Microsoft to follow. I'll explain.
For those unfamiliar with my prior Microsoft/Power BI note, within this note I detailed a demonstration that Microsoft provided at the Data Insights Summit regarding newly introduced predictive/smart suggestion capacity via its BI engine Power BI. And while the newly introduced capacity was just that, a base level function meant for building on down the road, it was still entry by Microsoft and Power BI into an arena that I think will define the space over the next handful of years. I was excited and I was optimistic about the coming developments of the platform. After the Data Insights Summit I interviewed the CEO of what I consider to be the leader in Neuromorphic-IP licensing - Brainchip Holdings; neuromorphic capacity, again, is something I've been very promotional of as being the additive that can take Power BI to the "next level".
I want to reiterate Van Der Made's commentary, alluded to above, because I find it terribly relevant to both the news that Brainchip made public June 30, 2016 via a press release - that it has acquired a French-based vision technology company (with a focus on artificial intelligence and untrained pattern recognition) - and I find it terribly relevant to the potential future of Power BI at Microsoft:
" (via Peter Van Der Made, CTO and interim CEO of California-based Neuromorphic-IP licensing company Brainchip Holdings) If Microsoft isn't already investing heavily into solutions to take Power BI into "smart suggestion" territory that likely is only because it is looking to "buy" instead of develop; options are available from many base solutions. For instance, our SNAP is a market-ready solution, one that can literally be "plug and play" for Power BI, that can help provide an answer for one of the largest problems in financial data/analytics today - the ability to recognize patterns without having to be pre-programmed to recognize patterns. Put another way, while the majority of today's analytical engines are adequate for recognizing patterns once "deep-trained" to do so (we don't believe there is true "deep learning" in today's technology - but this is a subject for another day and another conversation) the problem is they quite literally have to be programmed for what to look for; this limits the spectrum of visibility for pattern recognition or the total scope of usage for today's typical analytical engine. SNAP, which again is one option (Microsoft has limitless optionality with its funding capacity), allows for a truly open architecture pattern recognition, especially when plugged into an engine such as Power BI (especially if correctly integrated with Cortana), in that our chip doesn't need to be programmed for frankly anything - it simply learns to recognize patterns on its own while it processes data and learns what is within normal parameters for the data itself and what is not. With other systems like deep learning, people have to know what to look for. With the SNAP technology the chip learns to recognize any patterns that are present in the data set, even if they are not obvious to a human observer. Imagine this value when presented via Power BI to a global financial firm or a global PR firm; even a global network of Twitter mentions could be monitored (to monitor sentiment) and presented via Power BI dashboards.
SNAP learns to recognize these patterns in seconds which could be helpful if Microsoft is serious about taking Cortana into coupled-platform areas with Power BI. Deep Learning solutions need to be trained with the patterns that the user wants to recognize; they have to be known. Training a deep learning system is a lengthy operation that can take days or weeks, depending on how complex the patterns are. Brainchip SNAP learns in real time to extract patterns and label them in the data set - something that Cortana is rumored to be working on in development for natural language query.
Today's "deep learning" or "cognitive" analytical engines are the equivalent of stargazing through a piece of rolled paper - you can see into space but with limited scope and with limited capacity. SNAP, we believe has already evidenced and can evidence on demand, is a way to view the stars using a high powered telescope and/or using a mapping engine that takes in the entirety of space. It's simply a better way to drive business outcomes in that it truly is a compliment to a fine instrument such as Power BI - an instrument that can present data in easy to understand ways for easy to execute actionable understandings. Microsoft has been active as of late in acquiring "SNAP-lite" technology so we're quite aware that it has much broader, much higher ceiling intentions for Power BI. We wish them luck."
First, I think it relevant to state that both myself and my small consulting firm have a huge focus on data and analytics and from this stems my background into Power BI and BI in general. I consider myself a Power BI power user and somebody who is intimately familiar with the platforms "ins and outs"; I volunteer this in that I hope this provides some background into why I think I should be able to suggest product expansion (from a user's standpoint). Power BI is the primary BI platform that our firm uses and it's incredibly powerful when partnered with a data platform like, say, Kamakura Corporation's "KRIS". KRIS, it should be noted, is a bit of a BI-lite platform in itself in that KRIS has data-visual creation as a feature of its engine (in an effort to make the data it's presenting smarter in suggestion and more predictive; smart suggestions and predictive ability is a theme of this update note). All told, I want Power BI to succeed from a selfish standpoint of wanting to be successful myself.
I also think it relevant to say that we use the combination of Power BI and KRIS (with a dash of Alteryx) to create what we think is a comprehensive, predictive data/analytics platform in deployment - something that I think Power BI could someday offer enterprise clients from within Power BI without outside integration of ancillary platforms. Put simply, I think that with evolution Power BI can eventually remove the middle man (the manual labor in this case) in pulling data, refining data (it already is to some extent with its close, almost in-house integration with Alteryx), and most importantly into providing real-time, smart suggestion, predictive insights into the data itself. Providing insights (and the hope is comprehensive, predictive insights) is literally the entire purpose of Power BI as a platform; because as Microsoft/Qlik/Tableau/Oracle/et al have found out "providing insights" into data is a hugely lucrative and hugely profitable business. I exampled in my prior Power BI update note how this has become a business more and more skewed to Microsoft and Power BI's dominance. To state the obvious, this has been good for Microsoft shareholders.
Now, regarding the next evolution in "providing insights" - the so called "smart suggestions" I keep coming back to and this being the Holy Grail of enterprise focused BI platforms - of course Power BI would simply be the analytics engine in the equation and not the actual data source (as that isn't what it's built for). But I think, as I've discussed with members of the Power BI team directly, that Power BI can eventually be the platform that makes smart suggestions to its users rather than simply making data higher in visibility and understanding. I also think Power BI would provide the more important, more valuable part of the "smart suggestions" equation. Being a data source is one thing, providing smart suggestions without additional labor being required is another.
And again, this matters to Microsoft investors. I think the future of Power BI - what I believe is a growth engine for Microsoft going forward in that I know that the Power BI team at Microsoft has much, much larger aspirations and intentions for the platform from a holistic integrations standpoint (with the broader product/capacity portfolio at Microsoft) - can be to the BI arena what salesforce.com was/is to the CRM arena. I really do. I also know from speaking with Power BI team members Jono Luk and Lukasz Pawlowski that the future of Power BI involves the technology being omnipresent across all things Microsoft. This includes full blown Cortana integration, the ability to deploy Power BI via Microsoft's app development and servicing capacity, etc. Power BI is a real focus for the future at Microsoft and investors should track its development because I think it will contribute to equity pricing as much as Hololens, the evolution of the productivity suite, and/or any other technology or product in development.
This is why I think Microsoft's pace of development into this next evolution is so important. I think Microsoft realizes this as well given its recent coziness with Alteryx - a data refining/data workflow optimization engine that is essentially a part of Microsoft at this point (it might as well be as it's so closely integrated to Power BI). Alteryx has allowed what I would call "base level" suggestion capacity, via the Alteryx platform (meaning directly within the Alteryx platform) during pre-Power BI data upload processes, to the Power BI dynamic. What is the Power BI "dynamic"? The Power BI plus point-solution platform users can "build" by integrating other platforms into usage. Again, something I think Microsoft can optimize by simply combining best of breed point solution platforms directly within Power BI (as I've stated prior I think Alteryx eventually is taken out by Microsoft as well).
Microsoft was smart to get to Alteryx early in that Alteryx is currently providing a nice refinement engine to big data sets prior to drop-in to Power BI; this allows the refined analytics to be more easily presented and to be more of a value-add to users. But wouldn't it be an upgrade if Alteryx could simply sort through and refine the big data prior to upload without needing trim for specifics? Or put another way, and to steal a metaphor from Van Der Made above, wouldn't it be nice if the range of vision for the telescope (described by Van Der Made in the excerpt above) didn't need to be focused but could be kept as wide as possible? Certainly Microsoft competitors Dell, Oracle, and SAP have come to this realization - that the status quo of BI isn't the future of BI and that the status quo of BI won't implement much of a "switching cost" for users. The point to be made here is that this is a pretty well known fact - everybody in BI is chasing the Holy Grail of smart suggestions and predictive analytics.
This is where Brainchip comes in (the benefits of integrating Brainchip's SNAP technology were generally described by Van Der Made in the excerpt in the introduction of this note) - and where Alteryx would be truly helpful as well. Brainchip, which is just one name I'd think Microsoft would do well to consider (but I also think Brainchip's "SNAP" technology is unique and that Brainchip is the leader in neuromorphic-IP), becomes even more attractive to Microsoft with its recent acquisition of Spikenet Technology - an artificial intelligence/computer vision technology company based in France. Spikenet has already deployed its technology - called SNVision - into live, real-time data/analytics environments such as Airports, Casinos, and Manufacturing plants. Spikenet also calls the Bordeaux Airport, the French Interior Ministry, the Shanghai Police, and numerous Las Vegas casinos customers; meaningful customers which can provide a baseline level of concept prove out for Microsoft in due diligence. SNVision as a technology can most easily be described as a live-stimulus, real-time pattern recognition technology that can perform a range of both simple and complex tasks. Again, pattern recognition is huge into data and analysis and into the formation of smart suggestions. Pattern recognition, or the ability to recognize data points outside of a normal pattern (also known as outliers), is all SNVision does. By the way, Spikenet was able to mature its business to this point without having Brainchip's neuromorphic SNAP integration.
With SNAP integration, and this is my guess as to why Brainchip decided to make the Spikenet acquisition, I expect SNVision and SNAP to have expanded total neuromorphic capacity. Both should be highly complementary of each other from what I've read about SNVision and what I know about SNAP. I think that Spikenet can bring a very Datawatch-like pattern recognition to the visual mapping within Power BI (Datawatch is currently a close integration partner with Dell's BI solution) and I always thought that SNAP itself would bring ample untrained pattern recognition and continuous learning capacity to Power BI. This, to me, is an immediate switching cost that Microsoft can overlay onto the platform. Put simply, if Power BI literally learns patterns being created from within my data sets and makes smart suggestions as a result - say financial data sets being uploaded for risk management purposes - how can I possibly switch to a BI platform that 1) doesn't and/or 2) hasn't been "learning" for as long? The answer: I can't.
And I think that's the key takeaway here. That Microsoft wouldn't just be taking a competitive step forward with Power BI but it would be building into its platform the same high switching cost that has made its productivity suite (Word, Excel, PowerPoint, etc.) so hard to leave. It just makes sense. I've said it before and I'll say it again in conclusion here, Microsoft's Power BI is the best in the space and from my experience in using other platforms it's some distance away from its next closest competitor. That said, in my opinion, the first platform to achieve smart suggestions will be the ultimate winner in this high stakes game. That's a game that Microsoft has money to spend on and that Microsoft should spend on. If it does, look out above for its equity price as Power BI could/should be the next great product within the productivity suite.
Good luck everybody.
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article.