Ping An Insurance Company (OTCPK:PIAIF) Unveiling Ping An Group’s FinTech Transformation Conference Call July 14, 2017 2:00 AM ET
James Garner - Chief Strategist & Head of IR
Jessica Tan - Group EVP of Ping An Insurance
Ericson Chan - CEO of Ping An Tech
Ricky Ou - Chief Product Officer
Katen Mistry - HSBC
M. W. Kim - J. P. Morgan
Charles Zhou - Credit Suisse
Kelvin Chu - UBS
Good afternoon, ladies and gentlemen and welcome. For those of you that don’t know me, my name is James Garner. I joined Ping An Group right at the beginning of this year as Chief Strategist & Head of Investor Relations. I’ve been an outside observer of Ping An Group over eight years, but that said since joining I’ve come to understand many misunderstandings that the financial community have about our group in the way its run, which gradually we are correcting.
The thing however that shocked me the most positively since I joined was the scale and the sophistication of our technology. So I would actually go suffice to say, I believe Ping An Group is genuinely one of the most technologically advanced financial institutions in the world. So, today with great pleasure that we have with us, Jessica Tan and key members of her team, and they mean what, they have built on the technology front over the last six years, which goes far beyond Lufax which of course we are very proud of.
For those that don’t know, Jessica Tan is one of the most senior executives at Ping An Group and the key architecture, the technology build out which we are unveiling today. And I stress the world unveiling because I do not believe many investors and analysts here today or in our webcast really know what we have done, and I can say this with confidence that the simple reason that we’ve actually chosen not to externally communicate on it. So, what we’re going to talk about today is not some new strategy, it’s actually what the Group has been executing on for the last five or six years but we haven’t told you.
So in terms of today, Jessica is going to kick off the event with a 30-minute presentation. We take a short break because there is a lot to digest and then we’ll have presentations from Ericson Chan who is the CEO of Ping An Tech; and Ricky Ou, who is the Chief Products Officer of Ping An Tech.
I expect the presentations will last for about 1 hour 20 minutes in today. We’ll take another short break and then we’ll go straight into Q&A. So with my job done, without further ado, I’ll hand you over to our Tech Group, Jessica Tan, which is obviously the Regional Head today. So Jessica, thank you.
Hi, good afternoon everyone. My name is Jessica and welcome to our discussion today. As James mentioned, I am very proud to share with you what we've done over the past five years. This is something that we believe technology is a core backbone of Ping An Group. It's something that we power not just our core businesses perhaps the source of innovation for many of our new businesses of which Lufax and Good Doctor are one of many.
I'll start off with this, everyone, everything we do is focused on our mission. We want to be -- we are the largest diversified owner. We want to be the world leading personal financial services and provider. You've seen our 2016 financial performance, very solid, our revenue profits, and I think we're proud to say that for every customer, for owners we earned $312 profit. We both have few strokes as you seen from our 2016, in terms of access, in terms of scale of our financials.
So with customers, we've grown also very rapidly unlike some of the financial services institution where the internet users are fraction maybe a third of their total customer base, we've gone to inverse. You see that our internet user base is almost three times of our financial services customers and many of which are on that. And this is a transformation, as James mentioned that we're consistently building over five years. Of course and you know our ranking and I'll not dwell much on it, but how do we do this? How do we think about doing this? And this is something that I want to share with you today.
We really think about this in two focus, I think you've always seen that our core businesses, the core traditional businesses we have our banking, insurance, asset management and then over the past five years we really created a new sets of pillars what upon new businesses over four ecosystems. And the way we think about our four ecosystems is as following. We actually went and we believe the financial services, we're really changing the way that’s been provided to people over the next five to ten years.
We believe that financial services should not be bought as a product. We believe financial services should actually be integrated with everyday aspects of our life. So actually look in very systematic fashion at each of the industries that we think are most relevant, not just from a GDP standpoint, but also from an individual standpoint. So, as you can see here, we think that four sectors are most relevant for us and for our customers. Financial services itself, real estate, auto and health, the thing that is important to the individual themselves and to the economy.
And therefore five to six years ago, we've created everything around these ecosystems. And here that is what I want to share with you, really everyone knows about our left-hand side the story in the house, which is really our core businesses, you know our stellar track record in each of our businesses. But I want to share more about on the right-hand side of our businesses today and underlying technology backbone that powers the entire group.
On the right-hand side, this is how we think about this. What I mentioned is really about ecosystem. In each of these four ecosystem, our thinking is really, how do we take the technology and how do we invest financial services into each of these ecosystems so that we can make the life of the consumers easier. So they can work together with the rest of the providers in this ecosystem.
So, I'll use financial services ecosystem as a start today just, just to everyone understand how we think about this. So traditional financial services institution are traditionally very asset heavy, and so we think about really financial services in a very simplifying manner as one in which intermediates between assets and funds, and we think about this different business models, right, traditional businesses, essentially they do the business themselves. So they will actually have capital themselves to intermediate between the two providers.
Lufax for us was the very first company that we incubated to do a marketplace and the idea as everyone know, many of you are familiar with our Lufax is the marketplace whereby we try to get buyers and sellers of access on a platform, so that is the marketplace they can sell much more efficiently. And we as a platform provider, used our technology, used our risk, so we have our KYC and our KYP to better assess the needs of the customer and the needs of the product providers and to the match. And that’s really where the open marketplace is.
But not many know that beyond Lufax, over the years was incubated yet another company, we call OneConnect. And we think this is an even newer evolution of the model, it’s an even asset like it's a pure technology FinTech open platform model whereby we really served about 200 banks and 2000 non-bank financial institutions and purely using our technology. And therefore us is a virtue, is a virtue whereby we created a business purely based on the strength of our technology, and we believe just as how technology to help our core business grown very successfully.
We think that we can also offer to the rest of the 2,000 over institutions in the financial services ecosystem and help them with their business better. So Lufax everyone is familiar with, everyone might ask me complicated legal and organization structure of Lufax, but in an essence it’s really about an online emerging platform, right. So, you’ll see a very much what our Lufax colleagues will mention, know your customer and know your product. And in the essence that's what it does, right, which we spent a lot time investing over the past six years, technology about better assessing customer needs, the retail institutional investors and the products. What kind of products, what type of risk rating they into do matching? This is a heart what Lufax does?
Something little bit less known about Lufax, actually last month along with our Chairman, I was in Nanning right, a smaller, quite small city within China. But one of the innovations that we've taken now and a quite interesting concept is actually helping the government manage the asset liability. So, if you look in the local Chinese press, you'll see that the entire Nanning-government finances and over 200 companies which are state-owned within Nanning, their asset liability was actually run on our cloud platform. This is a something that our Lufax constantly reinventing itself.
We want to use the same thing that we’ve been doing on the online emerging market dimension and we will like to try it using that to help the government manage their finances. So it was really cool. We will get the facility of Nanning. We actually have the first online whereby the government actually issued one of the best online was in auction, was a blind auction. We were able to help them reduce the financing cost by 14 bps even that metro section, and we’re really proud of this new invention, and we hope that we can continue to do so for other parts of the government.
And once again move onto OneConnect. This is perhaps we’ve publicized much less. This is some -- the idea was really simple, we want to take what we have been using over the past 30 years technology that we helped our own core business and we want to do it the entire financial service ecosystem. As you can see the financial services ecosystem is now back over RMB200 trillion in assets and we are only 6 to 7.5 trillion assets. So we want to do the whole market, so we started the banks. Bank that's owned has about 87% of the assets and profits in the whole of the China financial services market, and then we created something for bank connect three years ago.
And we started all this with just one bank, little bank called Taizhou Bank and they really tried to help it. And then today we have about 200 over banks and 2,000 non-banks using the platform. And the idea is really simple, we created this whole FinTech layer where we connect to each of these banks core systems platform and then on top of it, they can run -- we run the white label apps for them, we run their SME loans. They will go to our platform where we help them assess the risk and they can do interbank trading. So for example last year loan, we had a RMB1 trillion in interbank trading on this platform.
Right now another example we have Bank of Shanghai which is very large bank. I think it's one of the top three banks in Shanghai we have a RMB1 trillion in assets. Their entire direct banking mobile runs on our platform. They will actually pay us where we will run and acquire customers and we'll do promotions and everything for them. So, this is something that we’re very excited because we find that you know if technology can help our own businesses, we can also help the remaining of their financial services sector.
So this is a little bit what we offer. We usually don’t share with public and I think James said that with really that at IR, so he specially organizes as far as for us. We’ve been mostly talking about this to our bank and non-bank institutions, but essentially where we help them to do -- we help to them to do their retail banking. As you know a lot of the -- there's about 600 small medium banks in China which constitute about 40% of the RMB200 trillion in assets. And a lot of them are in little cities of provinces they may not have access to a lot of the technology or the talents that they have.
So and in retail given the entire banking pressure they are on, we help them basically do the entire digital banking, right. So they can even pay or they don’t have to change any of their core banking platform. We can tailor the structured products for them. We help them develop their own online loan products and all of you these just for a fee where they will have to invest heavily into their own systems. They can be online within one to two months. We’ll also do their loans as you know SME lending is a huge sector and a huge source of growth in China. There is about 20 trillion in SME loans by these small medium banks in China every year.
And you need -- it's a very faced, very paper-insensitive where face-to-face high note of business, and we provide a platform is actually there some block-chain involved in it whereby we help them gather real-time information about the individual SMEs and can help them better access. And then we charge them a fee for it. And then also said, the asset into FY asset trading, the small medium banks every year has about RMB650 trillion in trading, but two-thirds of them is actually very inefficient because a lot of them to have very nationwide networks.
They are not able to go through brokers so there much higher fee, going through our platform can help them reduce their financing cost by 5 or 10 bps and they can do it directly with them. It’s a fully block-chain thing. We actually taken the smart contract concept and we build this where they can issue their own products on block-chain and then they can trade amongst themselves. So far there has been about RMB1.3 trillion on our block-chain asset platform and we’re really excited about that.
And the one other thing we've taken the one step further, last month, because we had this alliance about 224 banks with us. Last month, we tried something even more interesting. We said, look, we’re too slow at -- there is too many of -- I mean we're not enough to tailor everything for you. So we did an open API platform. So last month we've launched this, we've over 300 APIs. There are standard APIs that they can connect to and we make that open to them. We already have about 34 banks with their own technology team and their vendors who would develop apps and systems on our platform.
And we think that will even accelerate the progress because we're simply such as huge market. So we're very excited about this kind of new model. This epitomized what our entire group thinks about moving -- our Chairman will like us to move from one where we used some -- I don't know how to translate using capital to make money to something where equally we use technology in the Company. I think we see ourselves not just as a financial services institution but very much a technology company at heart. And we should hope that this is the beginning of how we're thinking about this.
This is an example. This is part of our OneConnect platform once the module which is very popular. We have many -- we have regular that many of our global banks including the lines of HSBC, Stand Chart even some of the large banks within China are using our credit bureau. The 85% of the 200 banks in 2000 FY as I mentioned, used our credit bureau. Just last month along we have 88 million inquires, a loan application that pass us through our credit bureau. And we’re very glad because this is something people are using it to basis on their actual loan decisions, right. Last year, we have accumulated about 100 million unique borrowers, loan applications coming through us. And we're glad that we're actually helping institutions and helping consumers in that way.
I want to talk a little bit then to our health ecosystem, something that we're very excited and very important for our consumers. And the way we think about ecosystem, I think in our previous Investor Day we shared before, we really think about in the three piece way. Like patient or consumer because you know secure, providers and of course who is paying for it. And the way we think about this is that we started all as an insurance company where our strength is really how do you manage health related costs, that's -- but in China as most of you know, the bulk of the health care costs is actually being borne by the social health insurance. So six years ago actually we started then offering our services to various individual governments on how to manage their social health imbursements.
Today, I think we've over 280 cities on the life and increasing by the day where we help them manage their social care -- social healthcare reimbursements. So Xiamen for example, that's one the first city that we did it for. In Xiamen the population is 6 million people and 10,000 doctors out. Every time the doctor in the hospital, whey they prescribe something, it'll run through by our engine whereby we determine what's the right level of reimbursement that we should, what type -- what's the right level of meditation that we should have with this particular patient, in his particular illness, and that's real time.
I think just alone on average for the city, we're able to help the government reduce about 14% of their costs. I think this is huge impact for them not just economically but socially, as you know with the ageing population in China, the affordability and sustainability of social health insurance is very important. We think this is -- this should be very helpful for them. And then beside on the social health insurance, on the patient side as many of you know, in China would know, one of the problems of the healthcare in China is about having easy access to primary care services, right. Even for flu, often our colleagues will have to take half a day just to go to the hospital for flu for their kit.
So, what we started doing is that we created two things, we think about having online and offline. So, online was really our Good Doctor which many of you know about, we're really happy how it's grown within three years, but we have over 150 million users. Every day, we have about 444,000 increase on our app. I think that provides -- it shows the need in China for affordable online easy primary care question, right. So, usually they ask simple questions before then they go to the right doctors.
Then in the offline, something perhaps as less known to the public, we have this One [indiscernible], right, where -- whereby we try to also upgrade the capabilities of the clinics because I think there're about a 1 million or so clinics all over in China. We try to have a platform just like what we've done for OneConnect, but we've then created something based on the standards you know the NV and other to whereby we have a technology platform and then we have the right sense about how you should manage your patient, your health records. What's the right level so that you can qualify for the social health insurance reimbursements to better manage the healthcare?
And this is something that we started I think about two years ago and we've about one and a half years ago, and we've about 50,000 clinics right now on our platform. And then everything on this is actually on the Pan An Health Cloud, we hope eventually that we'll be able to make an impact in the healthcare economy whereby individuals can have better access in the integrated health records to them. So everything there we have we connected to the health cloud. You as a consumer, if you're in the particular hospital you can actually us to request and consolidate your health records across these various providers.
So, this is an example of the social health insurance that we have and you see -- this is Xiamen, this is the first one I mentioned at where even the CRRC President at that time actually went to visit. It’s not only help save cost, but it’s also huge improvement in satisfaction, a great service and then we got a lot of violation. We are expanding that actually to not just include social health insurance. In some cities, we're doing a pilot whereby we try to integrate different parts of the city services on to an app for the government. So not just social healthcare insurance, but we try to do some like their health records, their long-term insurance, public services, various things and we help the government integrate that into an easy app that the consumers can look into.
Lufax and Good Doctor, many of you know we’ve grown very rapidly, now about 1,000 full time doctors. We work with about 60,000 part-time doctors and every day we have about 430,000 queries. So, I have to keep switch this in Chinese and English, my brain is all fried up. But we think this is really helpful. So now, at first we started doing a free online inquiry, last year we actually started with card, with the health card, you can pay about USD100, right. And you get about 10 times, 12 times free inquiry and after that we’ll actually have referrals whereby we on prepaid network and you can get access. You get one hour merchant delivery, if you listen to some of the cities that we do. So we are building a membership around here.
This is the clinic one that I mentioned just about 50,000 of them. There is extra view again I apologize. And this is something that hopefully over the long term is going to be difficult, but we hope to build a good brand of affiliation of clinics that has good standards. Up then there is two, I’ll just cover very briefly before we take a short break. There is auto ecosystem and health, which we also don’t spend much time talking about. These are ecosystems that we started to gradually think about. In auto, as some of you may know, we’ve actually invested in Autohome. Autohome is the largest comp auto in China, it has about 76% market share in the market share in the market in China. That means every year there’s about 22 million people buying new cars, second hand cars, 76% of them does it through Autohome.
On the average you will see that a person will comes here, they’ll spend -- some of one who want to buy a car, we spend on average 80 days on our app or website looking for the right car that they want and then before they actually go to the preferred dealer through us to buy the car. So we think it is wonderful thing and one of the reason we invest in this because we think it has a lot of synergies for P&C insurance and of course our bank as you know we are the second largest car insurer in China. You should also look actually our Ping An Bank is the largest car financing company both to the dealers as far as the consumers.
So we want to do something for the car owners. So one of the things that we are gradually building, you see in the Autohome car release is that, not only for the customers we serve them, if you're trying to buy a car through Autohome, but also if you already have a car then through our P&C, we have app for good car driver. And then we are also then working with our ecosystem partners, offline to view as in platform so that we can better link the customers to the providers. So for example therefore that Autohome as announced that they are doing, last month Autohome launched this car dealers, there are 29,000 car dealers in China there so the bulk of the new cars. We work with them whereby through this they would get better access to the people who want to buy cars. We will also offer right the financing and insurance through the platform, both for the dealers as well as to the consumers.
We also have second hand cars, second hand cars as you know in China is still a very underpenetrated market, but it’s going to reverse like just that -- and then in the U.S. actually therefore second hand car sold to every new car. We think that’s a trend but it's a very fragmented market, but now we are also trying to do that so that we can have better access to it.
And then we also have these car components. This is a RMB400 billion business every year whereby in parts a lot of it increasingly is not just through the car manufacturers, but is actually through dealers repair shops as the 100,000 repair shops. And this is something that P&C because we worked with all these repair shops, we know exactly the parts, the labor hours of every make across these thousands repair shops, actually we’re not viewing an open platform whereby you can afford. So overtime, this is still in the working stages but this is something that we think, we have something very interesting in the auto ecosystem which is a very important part of the economy that overtime we hope to do the fine.
And then finally the housing, real estate, I think this is also very long-term play but clearly that’s very important for the economy and the consumer. And here our approach has really been -- we have two companies which is also we don’t publicize too much. We have a Ping An Haofang, which really is an online housing portal. And then we have an offline mortgage center which is called [indiscernible].
And the idea here is that, it’s a very fragmented market in China. But we want to work what we work best which is on financial services. And mortgage in China takes more than a month, I may have never bought a house, but at least my colleagues have done so you have to make multiple trips to the banks, to [indiscernible] to get it, right. And what we've done really we take how to really streamline the process, the risk assessment so that we can get it, our fastest now is four days. And so we work this with the top 100 developers, the top 100 real estate agency in China. And we try to make finance the easy with their customer and that’s how we’re thinking about it.
Again this is one of the early stages and we’ll welcome any comments as well but we think this will become a very important part going forward. So that’s really what I wanted to share at the beginning. It’s really about the right popular pillar. What I want to impress upon is that we are not just a financial services company as mentioned. We are actually very much a technology company and especially something that we will be doing consistently over the past five to six years. The way that we think about technology company is not just of our own consumption, we actually have a very much platform driven kind of thing where we will incubate them, we’ll put them into ecosystems and we want to share that with the rest of the ecosystems players.
We want to take a short break 10 minutes and then after which I wanted to invite my collogues Ericson and Ricky, who wish as of you the technology backbone the powers behind both the left and right pillars that I think would be quite interesting, actually a lot of things in our lag that we usually don’t share, but James has said that now we got to be out there to share. So, we're happy when -- there is something new for us, so hopefully we're enjoying the rest of the discussion.
Hi, everyone. Thank you for being here. I am Ericson Chan, Head of Ping An Technology. Ping An Technology is one of the subsidiaries of Ping An Group, one of the 28 companies. What we do? We provide IT solutions and services to the whole group and in addition to that we have hundreds of other corporate customers. Of which many of them are insurance company and banks. In beyond that we also have a very strong research and development institute. The focus is obviously is FinTech. And in the addition to FinTech, we focus a lot of our asset recently on artificial intelligence and I would like to share a little bit of that today with you.
Okay, you will have seen this chart from Jessica earlier. So, we the Orange Foundation is what we do and support everything for the entire Ping An Group and our customers. So how to make us so successful so far, I think there are four key elements that we have been doing. One is the rich scenarios or which of used case that we have, one is data that we have been able to leverage and the speed, speed of execution that is very, very paramount to us. And lastly as I mention is how we have been leveraging artificial intelligence. And many of them is developed in house recently.
And again you have seen this chart, I think this is their critical because this is our strategy, is Ping An Group strategy. So, you have seen the financial arms of ours on your left and then we have the four ecosystems, Jessica talked about. And then we have different technology models as integrated finance plus internet or internet plus finance or I call it life like a life tech, which is supporting different elements of the ecosystem. Off course housing, auto, healthcare these are the most critical decision in once life to make any financial decision.
And one why we are, we have so many different scenarios other than the ecosystem we built. We also have full license, we have full license of all financial products. So which give us the opportunity to serve our customers end-to-end. Okay, I will spend some time here to explain some of the things that Jessica mentioned earlier. What are the key things as Jessica mentioned at a beginning is that we have 137 million customers, financial customers. And then we have 376 million online users, that is very, very unusual. For most banks or insurance company, how do you acquire customer. You acquire customer by selling a financial product, it could be a top order insurance, it could be mortgage, it could be credit card.
You start with that. Once you get a product sold and you try to service them. When you get a chance to service them, you try to do cross sell. And then once you bring in the customer so you also try to migrate them, so called migration, make them online, right. And most of the banks today it'll be hovering around 30% to 40% of the online user base. Ours is completely different, we've -- we're using a different business model. Ours is 270%, it's fundamentally is different than all the other banks and insurance company.
Household, we first provide free service to address our customer pain point. This pain point is top of their daily lives, it could be auto, it could be housing, it could be any aspect of their daily life. So we try to use technology, of course without recent development of technology, that's not possible because of the cost because now we're leveraging technology we can provide services to everyone free of charge, no different than a lot of the startup today, or Facebook, the one that we're very used to today. We first engage the customer by servicing them before we start to talk about products and this is an example. You see the number, it's 270% of our financial customers are our online users today.
So, I'll use one of -- you see some of the screenshots, this is a -- I picked it from my phone. And this is one of the app that Jessica has mentioned before. It's called good car driver or Good Car Owner, Ping An Health sell to AVP. This is particular -- what this app does? This is a one stop shop for every single thing a car driver or car owner ever needs. It's a simple as okay where you'd -- where's your gas station to how do I do the maintenance of my car, if I need to do the maintenance, where can I get a repair shop, maintenance shop, if you want to read a little bit more about the auto news. It'll engage you, so like even a top gear-ish news bit, you can get it. And even -- one of the most popular functions that will be managing your traffic tickets, as in China it is an ordeal to manage your traffic ticket because sometimes if you register your address wrongly, even if you get a screen ticket you might not know, if you don't pay you get yourself in trouble. So this is an ordeal to manage this in China and this is a pain point we think that out.
So, what we do is all you need to do, you don't need to pay anything and you register your car and then as soon as if you do get a ticket, we will remind you. We will alert you when you need to pay. As soon as you need to pay and you recognize it, if you link it to your account then we'll settle the bill for you. So this is a one stop shop of everything you ever need as a car driver. Second, screenshot from the right, what is that, I don't know if you can read it from the back. This is the math actually is not too proud of it, it's about two months ago, when I was in a hurry, I needed to take my son to play badminton, I was late, so I got out from my home, and you'll see some of the red dot is when I turned the corner at it too fast. When I -- the orange dot is actually is a little rocket picture, that's obviously I was driving a little bit too fast. I was in a hurry.
So what it does, it's able to capture your driving behavior. Why do we do that? Once you capture driving behavior, so we can alert your you, hey, if you drive like this you're wasting gas, wasting gas is wasting money, so we are trying to engage you by helping -- there's no difference than many of us is wearing this kind of wearable device. So that you know a bit more about yourself, so we can manage everything a little bit better but of course by doing so we know you a lot better at the same time as a company so when you do the insurance renew, so we can rebate or engage you differently. So this is a very different way of trying to engage customer before we talk about product.
So, once we bring in a new customer, how do we -- because Ping An group is a huge organization. We have all different financials as a class offering and for example we have securities, we have trust, we have bank offering credit cards, mortgages. So if once you get one of the products, how can we engage to a different organization, and we have a very-very unique way that is called run even. So if like a -- how do you, it's almost like a teleportation mechanism. Teleportation between one company one app to another. You look at the screenshot at the back at the bottom, sometimes regardless which application you're starting with within the Ping An organization, all of them will have this style run even teleportation mechanism to tailor may suggest you what you might else, what else you might need.
So if your car -- have a car insurance with us at certain point in time where you think certain products like a credit card, it will fit your lifestyle we will suggest that. It will in one of tiles and also suggests other services that might address some of your daily pinpoints and add value to you. So by doing so, we can cross -- I don’t want to use the work but it’s cross-selling. We can use, do a lot more cross-selling between different products. So there's over 100 million customer of ours when they get on the app and they were able to use this mechanism teleport to another area and other company, another app of ours. So, we can engage them and then with more products and services.
And in addition and also you can imagine because of the account structure, it's not sometimes, it's not that simple, but referring you from one entity to another therefore we have another mechanism, it’s called master account. So once you tie your primary app with this master account then you can engage all other products from other app within the entire group, make it very seamless, painless, so financial products is painless, that is a new thing. So this is what we have been working on in the last couple of years. So in this we just talk about how we engage customer differently to bring them in and how we transfer from one entity, one asset class, one product to another. But regardless which product, which services and backend we try to digitize, automate everything possible, and we believe in technology and this is very fortunate of me to be part of this visionary organization, so we can invest and digitize everything.
So from acquisition to service to marketing product all the way, risk control -- risk is very important for our financial organizations. So regardless is due for 2C customer or even 2B customer, I don’t have any material but I can also talk about for 2B customer because of or AI model, we even put a lot of engine to find the motto, how do we manage risk better on our corporate customer portfolio. So with some of the engine because we’ve used all different data feed and also understand our customer better, we can also control the risk portfolio a lot earlier. We can have a early alert instead of three months, could be six months. And the cables of alert we can get accurately is four times before we have this cut model.
So we are digitizing and automating every end from acquisition and risk control. So that is the tough thing about these scenarios. The number of used case that we have is very rich. The same time once we have all these scenarios with so many customers able to use so many different of our apps then we can understand better and our data that we have been creating is very, very rich. Regardless it’s from the basic information of the customer, sometimes to education level, family, career of course investment and consumption and health.
So we have a very, very good understanding of our customer. Therefore we can provide tailor made service and products. And with this kind of data, we have a little army of scientists of 500 data scientists to try to bring insight and value ultimately to our customers. Speed, speed is very critical, speed to market, speed to provide services to our customers. One of the way, this is the technological way. We have started to built our cloud infrastructure, two to three years ago as of today we have already 70% of our own infrastructure is on the cloud. This is by far the highest in all financial institutions in China.
So what does it do, of course if your infrastructure is on the cloud, its more scalable, you can ramp up and ramp down quickly and even if you have a new product, a new system for those who know in the industry you need to get the production environment ready. Normally it would take days, probably weeks before you can get a whole new environment ready. With our home grown cloud technology, it would take us only 150 seconds. If you’ve decided, if our business decided to have a new financial product, we can start to click it 150 seconds later the infrastructure is ready for our coder to start coding.
In addition to doing coding, the entire SDLC flow, software development life cycle flow is automated. By doing so, we’re automating testing, regression testing. We’re automating deployment. This is 2.5 times faster than before. So as a result, 59% of our software release finish in 14 day from decided this is a product and service our customer need to the first customers who start using it, 89% of the software release we complete in 30 days this is very important for us. For this obviously is a very, very fast deployment speak to market for financial institute. But we don’t benchmark ourselves against the financial institute. We benchmark ourselves with internet startup, this is the pace we want to deliver but the quality of service, security of service then we benchmark ourselves against financial institute.
Lastly as most exciting, everybody talks about today are artificial intelligence. We mentioned about cloud. We also focus a bit on the digital online to offline, office to online. But artificial intelligence is our research and development core today. We spend hefty amount on artificial intelligence. We have 20,000 R&D staff across the whole group. And many of them has a master and PhD degree, so they are smart, they are very smart. And I'm very, very proud because just this year we have 1,458 by now I'm sure the numbers are little bit higher patent applied. That is number one in all financial institutes in China. So this is to demonstrate how our smart guys turning out their work and generating value.
And for those of you unless interest in AI, we look at it in three categories perception, prediction and prescription. We engage in all three and spend a lot of our time in all three categories. So what are they? First one, perception, perception is basically this means when I see something, can I make some sense out of it, in this particular case. When I look at you do I you if you're you, very simply. I want to make sure that you're you. This is beyond what a FinTech normally as scoped, as you talked about straight official FinTech scope is, this is not in there is, not we do robo-advisory, we do block chain, we do through PA lending. We do mobile payment and wallet. But this is beyond that and this is still we believe is core to our business because this is important, its security and also its direct convenience for our customer.
So we're happy that in June we have hit world number one on accuracy. Our facial reorganization engine, according to LFW benchmark, in June we topped the world that is fantastic and we just had a celebration party last night. Thank you for the budget boss and then 300 million usage, this is only since the beginning of last year. Since the beginning of last year, we have 300 million times being used this model. But every time when you use it the engine will get more, the model would get more and more accurate. At the peak time this is a geek number, we were very proud of 30,000 faces we processes in a minute. And so what does it mean? Is very exciting for techie like myself but there will be more cases and Ricky will explain later on.
But I'll highlight with just a one case, how we have been using it. For example, one of our loan product, personal loan product in Pu Hui business is called iLoan. You have an app on your phone, you download it, you need to -- if you need to apply for loan, you take a few photos of your own ID and some of the supporting document and then you click it, it'll engage one of our agents through a video conference as free time a video conference. And we'll start it at a few asks. Why we're asking the questions? We will be able to identify if you're you. So surprisingly you'd think that will be quite easy, this is an important example. This is just last night after the celebration from Shenzhen coming to Hong Kong, Ricky and I was in the call.
And if you know how to cross the border between Shenzhen and Hong Kong, we get Hong Kong ID and then to the officer and then you look at -- as you look at ID, look at you, look at ID, a few times, it's just -- and then look at Ricky, look at -- ID look at Ricky, and three times, and then hey, can you take off your glasses. And okay, Ricky took off his glasses and look again, all right, okay then took another ID, look at me. All right, hi, Mr. Ricky, so, he -- this is -- the is the -- I think basically they -- the officer has mixed our two ID and just because he -- you can judge if he looks like me later on. Just because he took off his glasses, then the officer, okay, that his him then, that's me. So, this is why we take security very-very seriously, this is why we spend in the groups, invest so much into security like this. We want to not only safeguard our long portfolio also want to safeguard your money.
So, this is -- and as a result three minutes, it's in three minutes, the entire long process is completed and we're not only working on facial recognition just to know you're you, we want to know a bit more, because every face has 44 twitch, has 44 micro-expression, every micro-expression has five level movement, it'll generating five to 19 different emotions, this is what we've been researching, so by doing so we also can alert, hey if this customer is happy about our service, if they're not happy, it could create some alert for our agents, hey you better be careful, and we speak of this, that's a -- so this is not just about security, it's entire customer experience we're embedding into our noble processes that you might not be aware.
Another perception AI, voice recognition, voice recognition is very different, in fact you spent -- after we spent some time voice is even more technologically advanced than face. Order chart, you need to isolate the voice, isolate the voice and figure out if this is a voice is a recording or it's not a recording. And then you need to separate different -- because quite often a few people talk at a same time. When you separate different track, and then you draw out some different pattern as a result, now we've developed a voice recognition engine. What this engine with 95% accuracy and we're trying to make even higher. And then using different -- we've a fair bit of conversation that used to train our voice recognition engine.
So, how it works now just be like this, it's not asking you to read any particular script, like some of the vendor product would do that. We think that is not very natural for a customer. So by speaking like myself right now, 18 second of my speech I will able to identify if your patent, your particular patent. And then the next time when you call in 10 seconds of your conversation I can be certain if you are you. So by combining what we are doing is right now as combining facial recognition, voice recognition that will be a bullet proof to save your financial assets.
After perception knowing you are you what then as to do some projection so forecast. So this is one of the cases we -- we've have been working on in the last few months. And very proud it's not just about financial product this is about health. We have been working with Chongqing [ph] government. Because of all our financial health insurance, we have enough claims we know the data, of the Chongqing [ph] City. So we used the data to create a model. So that we know we predict what about in two weeks time, in four weeks time what kind of illness is coming. Why do we want to do that? Of course one is we wanted to make the public healthier and we are since in the health insurance business and you have healthier public as also its also better to our bottom line. So it's the win-win situation of a lot of the things we have been doing. And we would like to do more and we will do more.
And the last of the three categories is prescription. So now you know you are you, you can forecast and what are you going to do about it, using our artificial intelligence. This is a case we actually have this we have been using for a year already. So if you are one of our auto insurance customer if you are unfortunate, your car is damaged, you can use our app to take a few photos of your car, upload it, our backend engineering will able to determine hey it’s a Mercedes or its BMW. If it’s a Mercedes is E class or C class, if the E car C class and what your model we will figure that out and also we will figure if this have you done any Photoshop is it for real this photo and then we will able to determine by the engine itself.
The damage level of the panel, there will be four categories of each panel, 14 different areas. And then we will figure out if pay job, a body work or completely replace the panel. So why do we develop an engine like this? One of our customers in Southern China was during a typhoon, the car was damaged unfortunately. The customer took a few photo upload it before the typhoon left southern China the money is already in their account. This is why we do this kind of technology. So and just to mention this also we develop a lot of technology and Ricky will cover this a little bit more in that about different products. And we put most of them on our cloud platform. From the Ping An eyes infrastructure layer, the path layer so we have block chain as a service, we have some of the big data engine, even artificial intelligence. We would put them on the cloud.
First we put them on cloud so that order companies within Ping An Group can take advantage of. And they have. And after that we believe in sharing, we want to democratize, we want to democratize our technology to a much bigger customer base even other banks in China and outside China. So and therefore I think the demand of high quality better financial services will across the entire industry and it will ultimately benefit Ping An Group also and we have seen the benefit already. So this is how we have been leveraging the technology we’ve developed to create value for our customers.
All right, I think I will stop here and I will introduce Ricky to cover a bit more on some of the specific technology in-depth and he is our Chief Product Officer. Thank you everyone.
Hi, everyone. It’s good to be back to Hong Kong and give this presentation. I live in -- Ericson, I actually almost lived in Hong Kong and we commute to Shenzhen everyday to work so. I’m going to introduce the three key technologies today. One is facial recognition, the other is cloud and the third one is block chain. Okay, so first started with the facial recognition and thank you for Ericson keying up with the story from last night. So apparently me taking off my glasses will be exactly like Ericson, okay.
Now, seriously coming back to the facial recognition, actually, the facial recognition we just started about 20 years ago. It's nothing new but unfortunately back then they were using a technology just like pattern matching and what not, and the accuracy could never get over 90%. That’s why it's deemed to more as a research project, but never being useful. Now what changed? Two or three years ago, we started to apply AI, the deep learning technique. So to similarly, how your PC just like digital trial -- how to recognize your faces, what’s the difference, the subtle difference that sometimes even a human eye cannot see, right.
So, we've slowly, slowly, slowly be pushed to accuracy to about 99%. So just for the record, people’s eye, human’s eye at the best, best situation, the accuracy is about 97% but obviously not that we can tell, human is also being influenced by very different scenario being tired in a different light situation so and so forth. It’s not quite reliable. But computer, they are very reliable and once we achieved the 99% that become something that we can rely on even in the financial situation, right. So that’s where it comes from.
Now, and then Ericson already told you that, in June we actually hit the benchmark 99.8% which is top of the world. Besides the LFW, we‘ve actually gone through a similar verification, we’ve the police research institute number three. So why they’re important, they are actually the authority for online authentication. So literally for all the financial related transaction and why now we would need to go through that verification. So with that, we also achieve even a slightly higher score on that. So all-in-all it shows that you know from an accuracy standpoint, we are already one of the best in the world, right.
Now getting to the different scenario, I think there is a video here.
Okay thank you. So in two minutes I think you can appreciate all the different scenario not only on the financial scenario but all the different other scenario as well, right. So overall I mean 200 plus scenario that actually to me is understatement, I think we can expand to that. Okay, so why our tech is better then let's say the other leading Chinese technology company, right? First of all is the engine itself, I explained the engine is like your little brain, so you can image the model that we're constructing is like a highly smart student. So we have accuracy of 99.8, you can image we got a student with a IQ let's say a 150 to stop it. But that’s not enough, the second thing is you have to put them to a school that’s where the data, the second in terms of data. Data is like okay you go into a school, what you learn, what’s the material, how good are the different material, right.
Right data wise I think a lot of competitor, they also have enormous amount of data but yet there data is singular dimension, they got a total but they don’t know who you're right. But in Ping An because of our big data and also we have 30 million plus customer, they are not just regular customers, they are our financial customer. So what that means is we have no very multi-dimension of who they are, right. So with all those data combined together, we so intended model and to train that model and to train that model, and this is the second part of the training; the data that we have, the richness of the data that we have to train the model.
Now, they finally graduated as a medical doctor and then get to the third stage, which is the scenario. So even if you're graduate -- the smartest student, you graduate on top of the class, the best university but you are not ready to be a doctor yet, you need to be test it out in a different scenario. This is how third of average over the out customers, which is not only we have hundreds of financial scenario for them to test it on, but we also have security all the other scenario that just show on the video. So at all, after a while, you’re really ready to be adopted. Now, then you can imagine this doctor, you will be truly tested on you can stay, you’re smart, you have probably best training. Third it has the most scenarios to battle test it before we put them up. So that’s the core advantage in that.
Now, this is another layer, I think I can go a little bit deeper into what differentiated us from let's say BAT. Now, BAT is a really good on the online capturing, this is like in online world it's very easy to capture your identity and then what you do online and then how you transact online. This is very much proven. But when you move to the offline world, this is completely different. And you will see even BAT today struggle to getting into the offline scenario, because they cannot do; how do they identify; people walks around, how do you identify who they are, what they do and how do they transact. Now, they’ve all the advantage of, let's say BAT today have something in loss and this is where is Ping An Advantage comes in. With facial recognition, I just show you in all different scenario, we can easily identify who you are. And better yet facial recognition is something that I need your approval to get, because the minute you walk into my premise I have the right to take you photo, I can scan your face, I can call your folio out from our database and then I can tell who you are.
Second, we're working with the partner. Here I give you one example. There is Shenzhen Tong, which is similar to the Octopus Cards in Hong Kong. So Shenzhen Tong has about 40 million cards like Octopus Cards. Every day it has 10 million active user to use the card, but without the step number one, identity, all they seem in the system is that 10 million card going every day, everywhere but they don’t know who they are. So there is really no value in that amount of data. So the first time we held them is to identify. We moved them to a online situation using facial recognition and we can identify who you are. Once I can be lay view and that card instantly that data makes a lot of sense. Oh Ricky, Monday to Friday, this is your trouble pattern, I can infer this is where you live this is where you work. Over the weekend, you were spending in a different part of the world, you're going to a theme park, you're going to this so I can further infer you may have a family and things like that.
So all the certain, once the identification kicks in, all the data become much more rich and valuable. So the second part, we’re actually working with the partner to track and those offline scenario and data. And then three, that’s where the engage come in with Ping An Tech, one of the things that Jessica and Ericson mentioned, we actually aggregated all our financial product, including bank insurance all the product on to one platform. So we make it very easy for us to tailor whatever at that moment what you need in terms of product you will put it right in front of you. So imagine it’s like putting a 7/11 right next to you following at any movement whatever you need, I’ll give it to you at that point. That will dramatically increase the rate of the transaction condition. So find the three steps identifying, track truck and engage and we can apply that to many offline scenarios and helping not only us but our partners to create new ecosystem for the finance.
Now, we talked about the scenario earlier, so we’ve seen the book. So first of all, before we release the technology we first will battle test some within our group. So in the past year or so, this has been applied in over 70 scenario or close to 20 subsidiary. Ericson just mentioned earlier the [indiscernible] loan that we did. Previously on menu process will take up 24 hour to complete a loan process, but with the facial recognition the [indiscernible] and electronic signature we can narrow that whole pass down to three minute. And because of that, we will be able to fast up, so the scale of the loan application going from everyday in terms of tens of millions to close to a billion a day in terms of the loan process. So we can see that with the technology enablement, we actually closely helping our Ping An Group’s broadens the Company to great improved efficiency on that.
Outside the Ping An, as we just mentioned, so we are rapidly getting into the airport, the house bureaus and the social insurance in a way to help the other institute to increase the efficiency, just like in Shenzhen Airport everyday tens of thousands people walk by, how do you track and then take out a bad guy. So with our technology, we will be scanning all the places, coming into the airport and then match with a known database of bad guy and then hopefully we’ll pick them out before they get on to our plane.
Now, not only there, we’re actually working with partner. So this is one example, so we are working with [indiscernible] Auto Group is actually one of the largest commercial truck provider in China, in fact it’s enormous. I climbed up to one of their truck it's almost like two storey high and it’s kind of scary on top. So you can imagine someone is driving that car over 10 hour a day, if they get higher and accident happen that could create lot of damage. So we are working with them by using our expression technology to trying to detect whether the car driver is being tired or not. And if they show the tiredness, we can give them a warning to warn the driver and if that continues without any response, we can even remotely trying to slowdown the car or stop the car. So this is the next layer of technology we trying to work with our industrial partner to trying to improve their safety and efficiency.
Second part, our cloud. Cloud technology today, I think, everybody realize is it's not something that’s truly differentiated. In fact, you talk about BAT, they all have their cloud Amazon, Microsoft they all have their cloud. From a technology point of view, we are not really that much different. I would say we are just like at the same level as they are. But what makes us different is the top left bottom, the compliance, security and reliability. Because Ping An is a full licensed -- as with all the license that just come to lot of scrutiny so every month we’ve been closely monitored by [indiscernible], the Chinese Central Bank, all the compliance.
So with that everything that we do has to go through their checking and compliance. And we are the only car provider in China that was under that compliance watch every day. So because of that and that’s why early you see Jessica and Ericson mentioned that over 200 banks, they happily put their stuff onto our cloud because they are desperate and trust come in because they know they put their stuff onto our cloud is compliant. And none of the cloud provider in China can remotely claim that.
Now, infrastructure wise, of course, we are -- geo-redundancy, we have multi size and make sure that even in the case of disaster we can fully off-load the loads and to the other city. Beside our basic layer of what we call the ice layer and the pause layer, what differentiated us further is what we call the SaaS. Over the past 10-plus year, because the fact that we slowly move other our subsidiaries, IT system onto the cloud. We actually along the way develop some very verticalized solutions, such as insurance, security, fund all of those are now being able to -- those services are being able to be a provision on the cloud. So you can say we have some leverage verticalised SaaS offering. And really at the end of the day, the SaaS offering is also differentiated all the different car provider in there.
So as such, because earlier I mentioned that with all the trust and into our cloud, not only our entire group so up to about 80% of our Group’s IT really reside on Ping An Cloud. But outside of Ping An, we have over 20-plus banks, all the different cities and all the hosted onto our platform.
Lastly Blockchain, everybody heard about Blockchain. It's like the magic in the financial world. So let's take a look at what we’ve done with Blockchain. First of all is scenario, because Blockchain in a way is very, let's say, centralized technology, in move paradise on centralized identity to be a decentralized offering. So as such that it really can apply to a lot of different scenarios. We identify 12 to as a start, but actually it's going to be more than that. This is only in the financial scenario.
So I give one example, the Inter-bank Trading platform, Jessica mentioned earlier, we are related truly of transaction last year in this platform. So how does Blockchain help in this case? So previously, all those staff that we’ve done on Inter-bank transaction is done actually manually. You can imagine, it’s done by facts, a lot of that is done by stats. People will come into hey did you receive my money can you fax it over literally that was how it was being done before. But with Blockchain, we have the contract as long as the bank is in the bank, they have a electronic contract in mind, they can do a handshake, they can get the transaction happen, and this is a timeless seamlessly and all behind the backdrop. So as such once we apply the Blockchain everything will go smoothly electronically rather than manually on that.
Now the other part of that is the real time reconciliation. Before that without Blockchain everything again is have to be done manually. There's no synchronization, everyone is using a different ledger. And because of that again it takes hours, if not days, to complete a single reconciliation; so three plus two that's kind of the norm previously for the completion date. Now with Blockchain, Blockchain in addition to Blockchain, we actually put a unified measure and mass some of those system and platform that we put together, we were able to do everything again online and with a supervision platform underneath to govern the rules and regulation and make sure the transaction will take place completely and seamlessly. Again, we achieve a T-plus zero reconciliation time.
So again, what differentiated us from our competitor, we are not going to get into every single one of them. But you can see along with all the privacy, security, performance, principles and most important last thing is technicality. As you can see in my earlier example, the Inter-bank, the Unified ledger, we’re actually applying that technology pretty much random. So it’s not just something that's in the lab, but we're actually moving the real world. So with that that concluded my presentation. And I think we can take another five to ten minute break before we come back for the questions.
Unidentified Company Representative
Okay, great everybody. Welcome back. This segment now we're going to open it up to question and answers. I'll make one request. Given that we've got three of our secretaries here, it probably makes sense given we've got about 30 minutes left for Q&A to actually focus the questions on some of the technology the team has presented today. When you ask a question, please state your name and the Company represent and try and limit it to two questions. That's two questions not two questions with five subsections. Okay, so who'd like to kick off with the first question? Okay, Nancy.
Q - Unidentified Analyst
Okay, thanks James and thanks for the management from Ping An for the great presentation. And I do have two and two only questions. And one is that you just mentioned that you the largest Big Data platforms monitoring financial enterprises and also the largest cloud amongst any financial companies. Can you put that into context straight up versus some of the largest big tech companies, and can you give us the context that how you start against those companies? So that's my first question. And the second question is, we all heard a lot of very exciting things. So I want to ask something that's a little bit different. So what keeps you awake at night, and what are your biggest challenges?
I think the first question is about the cloud, comparing our cloud with the rest. First of all, we developed a cloud technology ourselves. So we're not leveraging any other vendor. And then we also, in addition to that, we make sure that it's like the financial industry grade. As unlike some of the other -- nothing the other is not secure, don't get it wrong. But we focus -- put a lot of effort on making sure our cloud technology is secure. And in addition to that, I don't know if you heard mentioned. We need to be complaint for all different regulators in China. When I say all regulators, probably I think we'll be quite unique is the [indiscernible] is a Central Bank, the banking regulator -- insurance regulator and securities regulator. And they all do have different, a little bit different set of rules. And we need to make sure they're complaint to all, unlike if you -- any other Internet-based cloud provider, they don't need to do this. So the level of stringency that we offer is quite different.
So we need to put some numbers into context, and I think you wanted it relative to that I think there's two aspects to look at it, I think one is just pure scale. And if you look at pure scale, let's say it is cloud by revenues from what I know our cloud revenues just high as 15% of the large public cloud provider, so that I think give some sense. But I think more importantly, as Ericson mentioned, both of the data and the cloud aspects. The reason we stressed is in the financial services sector. It's a very different nature. It took us -- we start building our own cloud five years ago. It takes a lot of work, I mean, we maybe look very easy. But when we first came to Ping An, we’re completely cloud less. We have 700 firewalls separating all our systems across 28 different companies. It took us really four years, four and half years actually to be exact, to move 70% of those one-by-one to a cloud infrastructure and we almost customized everything correctly.
We didn't start off wanting to build our cloud with the goal, because we evaluated and we didn’t think that the ones out with that public cloud fix on. So I think that's why it's first, a very different purpose. The other example I'll give in terms of data, it's not just -- I think ours is probably quite the size of ICBC in terms of scale of data. The Internet companies can actually have much more -- actually I don't know how much more. I think it's not just the amount of data, I think it's about the richness of the data.
One of the reasons that you see, we were not the only company that got the provisional credit bureau license. We were one of eight, Audi, Tencent, [indiscernible] credit bureau license. But if you look at how we’ve evolved over the past 2.5 years, since we started, many of the other Internet companies who got this, they use more consumer product when you see it on their approximately they use it for more everyday life. You have a score above a certain level, you can use the VIP channel of the airport in Beijing or you can apply for a visa faster; whereas, how is this actually used by 2,000 financial institutions or loan approval purposes.
I think this does not, by coincidence, because when you took data, it’s not enough. You need to have a full breadth of data and that you also need to have the fast track record in order to do the model. Just having a data without able to correlate who is good or a bad loan with a good or bad driver is useless, and I think that itself shows why we have both the track record and the richness of the data to do this.
Unidentified Company Representative
And the second part was what keeps you awake at night?
That’s right, too many things to do. Thank you. Many of these things take a long time, the Blockchain, for example. Lots of people talk about Blockchain. We went to U.S. we’ve talked to many people. We are not the first ones to do Blockchain, but many of them over three years. I mean, Ericson used to work for his better half in the past franchise he was there for 18 years. Not to say anything bad about, because I think it’s one of the early pioneers many of the global banks the Blockchain two years ago. But none has able to put it in realized applications. And I think what gives us a way is just seek of things we’re doing. It takes a lot of work to take something from the technology and actually use it.
And this one is exciting about, actually all three of us unfortunately are not Mainland Chinese, they travel every day to Shenzheng. I travel every week from Singapore to Shanghai and Shenzheng. But we’re really excited to be here, because it's very rare and I used to be at McKenzie, 13-years serving financial. It’s very rare where you get an opportunity with the breadth of businesses in an open recession to try these things. When we did that Blockchain we literally had four BUs. Now everyone talk about Blockchain to reduce reconciliation, nobody even do that. We talk to four companies I want to do this. We did it in four months. We were able to reduce 80% of the T+1 reconciliation all have problems real-time. And the Ping An is just what -- deal for Ping An is execution constantly just do it. On facial recognition and our celebration last night, the fact that we don’t have enough Chinese faces to do the model. We told Ping An Tech employees, those were the 2,000 employees were our first customers. We bought 100 cameras and we just ask people to sit there and take pictures all over their angle and do the first model. And it’s just really fun and I think that is a lot of things to do, a lot of work but that’s really exciting.
Unidentified Company Representative
Next question please. We’ll take one from the lady in white.
My name is JoAnne. May I ask the questions Cantonese or in English?
So the question is about Lufax IPO. Every time there will such question, that’s fine because every time just as we -- I think last time we mentioned this was probably in our shareholders’ meeting. We’re always actively preparing and we will look for the right time that is just our shareholders to vote. So unfortunately I don’t have a new answer. Thank you.
Unidentified Company Representative
Thanks. Actually our tech analyst will have two questions.
I am tech analyst looking, so two questions -- so one thing is I was just right I'm looking the global R&D spending of the tech company. So if you look you 7 billion number, actually if I rank it, I think globally ranked 14 or something. It's actually just behind some great Internet company, Amazon or Google or this company. It’s huge money, but just only set of the spending. But it looks the cost record of this company, we have roughly 10% or 15% of the revenue spend. So I think my question is that -- so I looked the difference between tech company and the financial company that it tech divisional financial company typically is a cost center but currently, -- but tech R&D of the tech companies revenue center to generate the new revenue for the Company. So how your Ping An Tech generating new revenue stream for the Ping An Group. I think that is the first question.
So first, I have to thank both our Chairman actually our Group’s CFO and CFO. So actually at the beginning they create thing on track as a P&L entity, even when I came first five years ago. It doesn’t have -- it's not cost center, it's not budget it has a separate P&L on its own. It's been taught on day one, you are going to survive on your own. It's not granted that people will give you the money you got to earn your money, earn your money in both internally and externally. And I think that has helped to keep the group technology in a little bit more market centric, they need to be the best because otherwise our company can chose so source from other people. So I think that has helped.
Now in terms of your question directly about how much -- how do we major the value that we invested out. I think there is really two ways we look at the value created all of the technology spend. First one is actually in our core businesses. If you look at our -- very across results, lots of analysts ask us how is that that you’ve been able to sustain a lower cost income ratio for P&C. Consistently over the year, before prices and regulations and after pricing regulations, it's not by accident. The way we manage our P&C been applying technology, we are able right now we have about 40% of our onsite investigation we can reach there within five minutes or 10 minutes, faster than a police car in China actually. You should call us if you ever get into trouble instead of calling police.
And we're able to do that because actually overtime we do a risky map with every year we have 14 million claims, so actually can predict where are the accident areas. We actually have this great dynamically shape every five to 10 minutes to parts of the times, stuff like that. So what I'm trying to say is that, obviously, the alone example. So I think we were the first online loan Puhui I remember one and half years ago. And today we are the largest consumer finance company, not by accident I think because of our technology. So I think that’s one way in which we’ve been able to create value, so that our core business can constantly stay at top of their game.
The second way which we’re trying to create value, I think this one takes a little bit while is the right sense, side of the house, I mentioned. We’re trying to take that as they look just serving our own as this is just not sufficient. We hope that each of these companies as good doctor, the one connect, the social health insurance, overtime becomes actually viable technology companies themselves, serve the needs during the market. I'm pleased to say, actually over the past six years, about three to four of our Internet companies have actually broke even and start making money. Now, a lot of doing the investments things, which used to do so just like I mean you cut your tech, Amazon, the famous story AWS they lost for 11 years. So we believe in the tech platform story, you need to invest at the beginning but once you gear, you get the scale. And so hopefully, over the next few years, you start to see value also from the right outside of house.
The second question is how to resolve the conference interview with your customer. So I mean that you are, as the Ping An Group is a financial institution, and the way you provide tax service is great example of financial cloud from technology. But your customer is also financial institution, instantly they are completing with Ping An as a group. So can you show some example of how you resolve the issue and become profitable business? Thank you.
So I think we do that at three levels. One is that legally, we create separate legal infrastructure so they are really, I mean -- Ping An is a very competitive group. We could be internally too so we’re all P&L driven, we have completely separate businesses. So for the one connect I mentioned the completely separate business and they treat Ping An bank like one of the 200 banks that they serve. At the security technical level each one of these, because remember even without serving the external banks and actually internally we have to manage also 28 companies differently, each one of them actually has a different regulator. We can't really share stuff like that.
So over the years, we built our infrastructure the way we segregate our data even our cloud. One of the unique things about cloud is that we are able to, our VPCs, virtual private networks is such that we really segregate. But in some cases, we physically segregate where we able to dynamically do that as well. And we’ve been doing it over the past 20 years at our own Company. So at the beginning of course, there is certain in this sector. I mean I can tell you I do that and so people may not trust you. So it took a while. Actually on its OneConnect started three years ago. The first one took a while but after that people believe us more and more. I mean the Bank of Shanghai, example we mentioned, now has been running live for nine months. And all their customers that were we will not comprise that thing. I mean ultimately, because we are regulated entity, we’ll never compromise because the one time we do something like that where we use your customer information that we shouldn’t, I mean we’re dead. We’re dead not just in terms of our new business but our core businesses as well. So this is something we take very seriously.
If I may just add one more line, so some of the customers really just told us it is good enough for Ping An, that would be good enough for me. So there's -- because they believe that if we can serve Ping An well in our scale, speed and security, they are more than glad to join.
Unidentified Company Representative
Right next question. The gentlemen over there…
Hi, it's Katen Mistry from HSBC. I’ve got a couple of questions. The first one is, just philosophically, why did you develop the technologies in-house, you could have gone externally. And then the second question is, given the business is very P&L focused, I’ve got a good feel of effectively the service provision to then feed into product provision. Could you go and give us an idea of how management has incentivized to encourage the behavior to benefit the rest of the group?
I’ll start with the second question, so how do you incentivize management…
How is your, the technology stream of management incentivize to effectively provide the services and then cross sell into the rest of the group. Because expected service initially is free. Thank you.
Okay, first of all, we have the scale. And many -- the customer base look fortunate for us, the customer base that we have is probably bigger than a lot of the tech company customer bases to begin with. But that is not the core reason. And some of the things that of course we spend like, for example, we talked about facial recognition and those in the market today, or when the time that when we first started, it just wasn't good enough for us and then we have tested the accuracy into a probability. And then Jessica also mentioned, it does not have a Chinese face model, some of the international supplier. So therefore, and we start to develop and once we gain scale ourselves and gain capability, actually we can do a better control of the quality and the direction of the technology capability of that products that we’ve been developing. And then it’s like a positive cycle and then we see the advantage then we just continue to do more. Not necessary that we are doing everything in house but a lot of it is a bit more than normal financial institute for sure.
Sometimes we also collaborate, collaborate with other higher universities, sometimes for example, some of latest leading edge technology. We’ll be working with MIT universities like that, so to help each other. I just met with this morning. In fact, I was like, MIT professor in Hong Kong. So we'll figure out what else we can co-develop. So it's not pure together but we, case-by-case basis, look into it.
And I think just on the first one just to emphasize, a lot about making this all of have that technology background. Often this is not about technology, it's about how you make it work that’s where the secret sauce is. The phase one is very practical, not just they cannot recognize agency, there's some local vendors too. But we need -- it's not accuracy rate, a lot of people don’t advertise forced positive rates. That's very important for financial systems. I cannot wrongly identify that that's the right person. And we need to be certain about less than 0.0001% and most people don't have that need, so we have to do it ourselves. On your second question about how do we incentivize, it's a multifaceted way of incentivizing because we're asking our tech people to do a couple of things.
So the three things we try to measure then on. I think one is stability and security that's at the heart of everything. This is -- I mean, you dive we ask and tell them, it doesn't matter how many facial recognition you do number one but we have a big security or stability there. And so that’s why I think Ricky mentioned this, the faster recovery every year, we do this, we imagine the entire data center got burst or something and we have to recover everything on the other side. When I first came five years ago, I think we were at 20 something hours to recover, we are now at 4.3 hours last year, we haven't done one this year, we should be in Q3, every year we do that. So I think that's number one. Number two is then about service. So we actually have a very complicated back-to-back anonymous satisfaction level, not just the usual service SLAs that most banks, global banks do, but like satisfaction and then for all the BUs, et cetera. So we make sure that what we’re really doing is what they want.
And then thirdly to incentivize some innovation, we have stock options and incentives just like any other Internet company. So actually our tech employees are one of the very rare exceptions. They get actually options in some of the business which we incubate, a lot of businesses are incubated from them we want them to say, look it doesn't matter where you incubate it, but it's like you can give birth but you don't need to grow. You didn't be the one to raise that somebody else can do it and you would still get the benefit so that's how we incentivize.
Unidentified Company Representative
Next question M W.
M W Kim
Thank you, James. This is M. W. Kim from JPMorgan. I have two questions one is about allotment for your AI development. So my understanding is, there's a two way of the AI development as a financial institution perhaps analyzing the data in very speed way only about the customer finance information what the numeric data would be one way. The other way is integrating the images and sound, should be another one, IBM Watson should be a good example on that. So I want to know as a financer that the FinTech company. So what would be the real roadmap for the AI for Ping An? The next question is about whether you are really focusing on the forecasting the claim for the customer in the process of underwriting. So many insurance companies nowadays is preferring the customer forecasting the customers’ input data and then they use it in the underwriting and the screening and make their growth ratio better. Do you have this kind of underwrite process currently or are you planning to do that if not? Thank you.
So I'll take a shot at the first question first, the roadmap. Yes, in terms of the AI roadmap, we're going in two directions, one is going at the depth and the other one is the breadth. First of all is the depth. So as you -- so earlier we do facial recognition so that's just telling who you are. The next that we are going to do is facial analysis. So the things I mentioned earlier, micro-expression. So we'll be able to currently recognize 19 different micro-expressions on your faces and be able to infer about a nice day of emotions state, the level of being happy or anger, frustrated and so on so forth. So with that, we actively incorporate that into our [indiscernible] [1.47.24] loan process. So in a way to trying to see whether you are nervous, when you’re being asked a question, so we drill down further. So technically do you lie, that scenario.
And the other dimension we had to go with AI is the breadth. So earlier we mentioned we have facial recognition, we’re going to have sand, down the road we’re going to retina and also the blood, the vein recognition. Why, why do we do that, because for every single biometric, there is always a weakness, so for example, fingerprint, it used to be like the very bulletproof but now in Taobao maybe $10 you can buy a fingerprint mode that you can store about 60% of the fingerprint and same thing with facial I mean you can take some of the selfie recording and you can pull a lot of facial recognition system. Now, what we do in the future, we’re going to combine those such as facial and voice together and then you combine the two together, your level of safety would increase dramatically because it will be very difficult for you to fake both at the same time.
I’m not sure we fully answered your question I’ll try the second one. So I think in auto, we’re doing three areas where we are focusing on; one is in terms of rating, and the new customers the renewal; and then one is in terms of place and then one is in terms of servicing. And I’ll just give examples of how we think what we do. So on the rating we don’t just look at your past claims because frankly everyone has access to that. And in China, there are these -- it’s publicized who has claim not claim et cetera. Then you just have the same rating card, which is why you see in China most of the insurers compete at the same price range.
So what we do is actually look deeper into what, for example, what the bad driver -- Ericson in his good car driver at but we also correlate with actually other data, there is a richness of data. So one interesting thing you see depending on your occupation, I’m not going to tell you which occupation, you actually have five times higher chance of accident than others just like simple, because we have transfer record and we have things. We’re actually looking to stuff in terms of health. There is a high correlation between people with high blood pressure with drivers. I’m not sure how much compare to your correlation there is, but these are interesting claims frankly, because of the richness of things we’re going to use.
For claims we tend to do, as I said, the damage assessment. It's not just identifying automatically which part is being damaged and how much, because we have accumulated millions and millions of parts over the years, how much it cost. We are now 85% accurate in terms of predicting how much the price is, that’s why in the example that Ericson mentioned, we can give you the money so confidently because we’re quite sure. I mean, 85 is still not good enough, we are aiming for 90% at the end of the year and these are the sort of things we can do, servicing the same things. We have these little chat box. If you try here on our customer, you can try our online check, 93% of our online inquiries are being answered by robots that we train over the years. And so these are things that we really apply on.
Unidentified Company Representative
Next question, gentlemen over there…
Thanks for exciting presentation. I'm from Nomura, I am the insurance analyst. My question is regarding your tech application. Do you face some challenge from regulation, and what the key challenge? Thanks.
Can you give an example of what kind of regulation you think we might face challenge with?
Such are regionally some headwinds from the P2P like some new regulations so, et cetera. Thanks.
Actually in our previous presently this question was asked me as well. I think what you see in China in the past five years what called Internet financing is a bit different from I think what we're trying to say today which is intact. Ping An really is the P2P company is a best example. Actually institutions we're trying to do financial services, which is on the Internet so they are actually a long company. So I think which is why we see over the past year class, a lot of regulations have come up because these used to be completely on regulated entity. And then we realized that these companies are actually doing financial services, therefore, they should be subjected to some regulations. So I think that has been the context, which is a bit different from what we're talking about today. Today, we’re talking about technology itself that has existed for it that keeps us longer serving financial institutions. So we are not doing the business of the fun. We’re saying, how do we use technology to help institutions who have licenses, do their financial services better by doing their service better, by doing their sales better, by doing their risk control better. So I think it's not really impacted by any of the regulations that’s being discussed today.
Unidentified Company Representative
Thanks you. This is Charles from Credit Suisse. So two questions, first of all, it's about your -- so you talked about the voice print recognition, face recognition and also noticed you use the word used by Lufax and P&C. So can I ask you like, can you give us a number of sense how much you have already use those AI technology in your business. For example, may be 10% of your claims have been used this technology, or can you quantify this one. Because I think investors want to get a feeling whether it is something you…
How much more…
And also, how much you have already used now and also -- so like how much has been really put into the reality in presentation. And also besides Lufax and P&C, what other Ping An business or units can also use this maybe in future? So this is the first question. And the second question, I think, you probably already talked about this one. But compare with other big firms or whatever, can you may be talk about trends Ping An comparison and also the street challenges? Thank you.
I think on the first one, it depends by different technology since we’re at the first stages. So facial, for example, is very widely used already, I think about 10 of our companies use it, our securities -- 17, so almost all in terms of the calculation there. Our securities companies just for the account opening, credit card users, et cetera, et cetera. Voiceprint is something relatively new, we've been working with MIT for three years on this and it is now in the first phase of actual live testing. It's actually really difficult because we're trying to develop a single print for the voice. And normal person can only differentiate about 100 to 150 peoples’ different voices we have to train the machine to recognize a billion people. And we've been able to make the breakthrough actually earlier this year, so we just launched it. So we will be applying, I think this year in three areas in our call center. We have about 200 million calls every year.
So we’ll be using that -- if you're a customer you can try that now because a lot of self service -- our self service rate today is only 50% on the phone-IVR. That’s already much higher than most of the -- most financial institutions in China is about 30%. We expect to increase that to 80%, 85% because some of people stop in the self service because they forgot their pin number, they forgot their card number. So we're going to use the Voiceprint and do that, and we’re very confident we can do that because we've been recording these peoples’ voices for years. So I think that one is still relatively new, we'll start the call center, we're doing in the P&C and Lufax. And then next month you'll see we’ll actually simultaneously launch other financial institutions. So one of the things -- just so how serious we are on the right inside of our house. We are going to offer to other financial institutions even before the rest of our companies can use it. We want to give it a level playing field. Whoever wants to use it we socially incentivize our own companies to come and use it. So that’s the first question.
On the second question what's our relative strengths, weaknesses compared to some of the T&T companies. I'll start the weaknesses first, that’s obvious one is that they have much more online users and traffic than us after all we could start off that. So we now have 400 million Internet users, they have 800 million, 900 million. So we’re caught up but not yet. The second thing is just the frequency of the data, many of expected social chats, et cetera. They've lot more data than we do. And then third one, they have much higher valuation, so much better than us. So I think these are the three weaknesses I can think about. I think in terms of strength, no, over the past five years, we’ve also found unique strength, which is why we are quite confident here.
I think the first strength is actually on the traffic. We’ve realized that in financial services and health, it’s not like selling a commodity item, the face-to-face matches. Just Lufax, for example, 50% of people who try to open account actually it's about halfway online. They do it by us calling back or by our agents being there. And then once they get on, they are much better. The face-to-face does matter and China is so big, most of them as more as serving educators and some of the younger generation in tier one city. And that is huge for us. Our 1.4 million agents is wonderful, they can sell anything, [indiscernible], for example which is one of the companies we own for prosthetic skin care. And we are now their largest distribution channel because our guys can sell. And I think that's very unique for us.
The number two I think is in terms of data, we've realized that it's not just about having lost data. I mentioned earlier, it’s about having the richness and the accuracy of the data. We have 880 million people -- data all of which we know exactly who they are. We don’t have to guess, I mean, if you are a search engine actually a lot of the AI is been testing who you are, they try to guess whether you're a male or a female we don't have to guess, our customers told us. And I think that's the advantage of the data plus the track record, so we can build models. And then the third thing I think is about risk and security. We grow up that way and the way that we do our financial services product risk control and the way in our systems, the ability and even how we do with regulators.
In China, the regulation is very unique. It's not a nationwide regulation it has nationwide regulations plus local regulation. And you need to have presence in local regulators, because if customers complain they have to go to somebody. If you want Internet, which is why you see over the past few years is some of the T&T players have launched these Internet finance companies. They can’t do a lot of -- none of that has done car for example. They can’t do car insurance, because you need to have local guys give service. And I think we are able to do that. So I think these three strengths really make us quite unique and very interesting.
Unidentified Company Representative
Next question, gentleman…
Hi, this Kelvin Chu from UBS. Can you talk about, on the life insurance business, in addition to underwriting the efficiency what kind of technology should we expect in the next few years for the life business that will help Ping An Life to further differentiate from peers. My second question is simple one, so 20,000 of highly intelligent individuals, are they just on AI or technology overall? Thank you.
You want to answer the second question then?
A lot of it, making not only on AI, but is making new technologies. So it could be making changes automating, so these are driving change for the entire group that is what that 20,000 technologies is for. It's not like, not keeping the lights on if this is what you're talking about.
On that, I'll just elaborate. It’s in practice for five years ago when I first came, we only had 4,000 developers today we have 20,000. Not only that, the number that Ericson showed earlier, 60% of something we deliver in two weeks. Five years ago, it was like 70% in one month. So we've maximized up to 10 times over the five years in terms of what we can churn out and that’s just in five years, imagine what it would be next five years. So I think that is just scale and speed in which we recognize. To your first question about life insurance, I think there's really three areas it's not just -- it's coming but we’ve been actually consistently doing. I think despite underwriting one is basically our agency productivity.
You've seen and lots of analysts asked us at every road show. Why is it that we've been able to have a much higher productivity at a much larger scale? Like some of our competitors have close to our productivity but they have one-tenth of our scales. And this is really because 10 years ago already, our Chairman has said, they want to make each of a agents like James Bond, we've got to make that equipment technology. So that any of agents will come, they can sell any product. And not because we're so smart or anything, we have -- our training app, we've a wondering training app that we did this year. It's online training app whereby they can simultaneous -- just flawless craze in China last year. We actually have to report for our training, 1.4 million agents being trained on use that. We equipped their laptop or their iPad with something is a remote -- we had this four years ago already whereby over 3G and have a video connection with our investment specialist in our centralized contact center. So that with their customer they can appear professional that's how they can sell.
So I think that's one area we're consistently working at and you can expect us to continue doing so in the future. And then the third area is servicing. I think one thing that you'll see is that we're going to not only have the chat box, I mentioned and improving our IVR and stuff, but we're going to radically change how our counter actually shop that. We still have a few thousand life insurance servicing full on the counter, which is actually highly inefficient, because the amount of time that people come to use. So we're going to actually radically change that and make it much more automated and self serviced.
Unidentified Company Representative
So we'll take a final question and then Jessica will wrap up. Lady, in the front.
I want to know which part you invested the most, and what's your focus on the next step incubation?
A - Unidentified Company Representative
Finance and healthcare are our two focuses. So in terms of these two areas, that's what we have invested in for the past five years and we will continue to do in the next five years. And we haven't done many areas in the healthcare industry, such as the pricing reform. But pricing reform is only one part of it. We hope we can -- so we have the City OneConnect and we want to develop more. So in these two ecosystems, you can see we will go very in-depth, so just the OneConnect I mentioned. We have over 200 banks, and over 2,000 financial institutions. But their product is only -- average product is only 1.4, so we can go much deeper than that. So in the next five years, we have a lot of things to do.
And in the other ecosystem, such as the auto and the real estate, we will take a gradual long time to develop the areas and the Autohome we acquired actually is only one year. You can see Autohome stock, the share price has increased 50% since we joined. And that also brings us a benefit and because P&C and the Ping An Bank are offline branches and outlets, but now you can see we can develop a lot. We have a huge room to develop the auto ecosystem and also really state is also a long-term plan.
That’s it for today’s event. Thank you very much for your attendance. I’m not sure if you want to say a few final words.
Sure. Firstly, I just want to thank everyone, particularly for the tech analyst group joined us. I hope you would come to us more often. We would love you to bring us new ideas. It’s a whole different world there. And then I think all we want you to takeaway today is that I think don’t just look at us from the left side of house, we’re very impressive. But hopefully, on the right side of the house, we’ll equally bring the value and then helpfully more will come. Thanks very much.