Distilled Identity

I'm Dave Shrier, the CEO and founder of Distilled Analytics. I started the company two years ago, together with Professor Alex Pentland at MIT and Professor Alex Lipton at the EPFL, because we saw that 3 and 1/2 billion people were being excluded from the financial system globally. And in particular, that digital identity holds the key to unlocking this financial inclusion.

I'd been really interested in this question for years of, how can technology solve big problems in the world? It is what drew me to MIT a few years ago to work full time helping run a research group. And then more recently, it's what led me to spin out some of the technology from that group into a startup company, Distilled Analytics.

Social physics is a new science that was invented at MIT by Professor Alex Pentland, which forms the heart of our technology in Distilled Analytics. It's how we are addressing this issue that there are 3 and 1/2 billion people who are insufficiently identified, who don't have good enough credit scores or any credit score at all. In some cases, they don't even have a legal identity.

Using this new AI-driven data science that's called social physics, we're now able to identify people, form better credit models, predict what they're going to do with their money, and help financial institutions then lend to them with confidence. Help governments identify citizens properly and then deliver services to them.

Along the way, we're solving problems like the $330 billion false decline problem. So it turns out, that if you want to use a credit card-- let's say, in the US-- most of the time-- in fact, 99% of the time-- it works. If you try and use that credit card overseas, however, it's a completely different story. So for example, in South Africa, your approval rate for a valid transaction is only 27%. Think about that.

Three times out of four, you're a legitimate customer. You try to buy something, and the machine dings you because it does not have a good enough handle on your identity. The fraud systems close you out. Globally, this problem is $330 billion a year. And some of the largest credit card companies in the world believe that our technology, built around social physics and some other research that we found at University of Oxford and at Imperial College London, can actually solve this problem and deliver the same kind of assurance that your transaction is going to go through that you have here in the US to someone elsewhere in the world.

[MUSIC PLAYING]

DAVID SHRIER: Predictive Identity is the name that we've come up with for this combination of powerful new research-driven technologies that let us deliver on the promise of inclusion for 3 and 1/2 billion people. One of the core issues is that the way that we do credit scores today, or even just the way that we identify who is someone that can do business with a bank is based on really, really old technology.

So this concept of the physical ID-- it's got our name on it, our picture, it may have some other basic identifying information-- dates back to 450 B.C. King Artaxerxes in the Tigris-Euphrates Basin gave some cuneiform tablets to the prophet Nehemiah. And that identity document hasn't really been updated very much in the last 2000 years.

But now, thanks to AI technology that we've captured from MIT, the University of Oxford, and Imperial College London, we have a much better way of identifying people. It's much more accurate. It's much more difficult to forge or hack. And it has powerful predictive components, meaning we not only understand who someone is, but we also can figure out what they're going to do with their money.

This leads us to the idea of a positive credit score. So for example, there are some financial institutions in Brazil that are very excited to work with us because today, we've now got the ability to look at the 80% or so of the Brazilian market. I'm sorry. I've got to back that one up.

Today, we have the ability to help the 73% of the Brazilian market, that right now has no access to credit, get into the credit system through the concept of this positive credit score that we build on our Predictive Identity software as a platform. Identity is the core platform technology that we deliver in a software as a service fashion.

Once you've got this much more robust, much stronger identity, we can then build apps on that platform, like a credit app that is 30% to 50% better than existing credit scores that you might get from someone like Equifax or Experian. We can also build fraud prediction apps that not only tell you where fraud is happening today but give you a window three months into the future as to where fraud might happen.

One question that comes up sometimes is this question of user privacy. People say to us, well, if you're using things like mobile phone technology or payment card transaction technology, is that somehow not good from a personal privacy standpoint? We spend a lot of time thinking about privacy. In fact, our researchers were integral to the thinking that led to Europe's GDPR, which is viewed as the global gold standard for privacy and personal data protection.

We handle it in two ways. First and foremost, user consent. We're not forcing you to use this technology. But if you want better security on your bank accounts or your other personal information, if you want the ability to get a loan where you might not be able to get a loan before, then you can opt in to be considered by this analytic engine and have that data used to help you with a healthier and better financial future.

The other way we handle privacy is with aggregation. So for some of our models, particularly around things like fraud and crime, we use aggregated information. So it's not one person, but it's an entire group of people, hundreds of people that are all pooled together.

It turns out, it works out to be about Zip plus 4 in terms of its resolution. That's considered to be good enough in terms of protecting personal privacy. So we can't really identify the individual, but still find enough resolution that our models are able to generate useful predictions.

We've got a powerful team that we've brought together to solve this problem of building better identity software for the world that then can deliver better credit and fraud systems to include 3 and 1/2 billion people. Professor Alex Pentland is my co-founder. He's founded over a dozen businesses out of research at MIT and is someone who is viewed as a global leader on big data, artificial intelligence, and digital privacy. He currently advises the board of AT&T, the UN Secretary General, the European Union, and others on questions like how big data and AI can be used to better the lot of humanity.

You'll note that I've got my UN Sustainable Development Goals pin on for this interview. Sandy is very involved with the use of big data to measure the SDGs. I'll also note that SDG 16.9 is the identity goal. The UN Secretary General has said that everyone on the planet needs to have an identity by 2030, and technology like ours can help reach that goal.

Our other team members include Alex Lipton, who wrote The Oxford Handbook of Credit Derivatives, and is viewed as one of the foremost authorities on quantitative analytics. My Chief Analytics Officer, Tom Fox, worked with Alex previously on Wall Street and has run multi-hundred person Quant groups but still rolls up his sleeve and codes in R and Python.

My Head of Engineering, Ellis Wong, ran engineering for Carrier IQ, a company that worked with all of the major telecom carriers and now is owned by Nielsen. It's how Nielsen measures mobile. Ellis previously has built an engineering team from scratch into a global leader in advanced software, and he's eager to do it again.

[MUSIC PLAYING]

DAVID SHRIER: This isn't my first startup. This is going to be the eighth company that I'm helping to lead in a C-level role, or the 12th if you include some board roles. So it's exciting nonetheless because this is the first time that I'm really getting to tap into the power of an institution like MIT to bring it to bear on a problem like global financial inclusion.

In the past, I've done things like helped Fortune 1000 companies build substantial divisions, such as the Enterprise Risk Management and Compliance Division of Dun & Bradstreet, and deployed capital as a venture capitalist. But I'm really excited now to be working at Distilled Analytics, a company that I started in part because I was inspired by my students at MIT. I created MIT's FinTech class, which was the first graduate FinTech class in North America. And eventually, when we took it online and expanded our portfolio to include University of Oxford, we've gone into over 120 countries inspiring FinTech entrepreneurs.

I looked at all these great students who were starting companies, and I felt a little sheepish that I was just opining from the sidelines instead of diving into the fray myself. And so that led me, in part, to start Distilled Analytics.

So the research behind Distilled Analytics dates back decades from work that Professor Pentland and his group have been conducting around the world. I'm trying to ask questions like, how can we use new technologies to bring more people into the system? What's interesting is that the future is the Global South.

If you look at the population today and the growth of economies worldwide, the biggest growth rates are going to come in the Global South rather than the Global North. This means you're going to have people who are more poor and more mobile enabled-- they're mobile first-- who are dealing with the financial system. The financial system today is not really set up to handle that.

Today's financial system is constructed around 10 million customers with an average balance of maybe $1,000. It has no way of handling a billion customers with an average balance of $10. But it turns out that if you take some of this research that Professor Pentland had conducted around social physics and around using big data extracted from things like mobile phones, and you apply it to financial services, you can create scalable systems to bridge to the customer of the future that represents the next 50 years of financial services.

There are other companies that are trying to use, let's say, identity technology to better resolve who people are, companies like Socure or SecuredTouch. In the beginning, you had companies that eventually were acquired by people like Apple that did basic physiological biometrics. I'm going to take your fingerprint, and I'll scan it off of a device on the phone and use that for account security and identification. A couple of problems with that, not the least of which is the fingerprint security on the iPhone 7 was cracked within 24 hours of its release. The facial recognition software on the iPhone 10 was cracked within 48 hours of its release.

And it's worth noting that Professor Pentland invented the original facial recognition software, the Eigenfaces, back in the early 1990s. So that's old technology. He's since moved on to much newer generation technology.

Some companies have started looking at ideas around behavior. But they're still kind of anchored in this question of who you are rather than not only who you are, but also what you're going to do, and that's what we call Biometrics 3.0. So Biometrics 1.0 is fingerprint or facial scan. Biometrics 2.0 is basic behavioral recognition.

Biometrics 3.0 is a combination of techniques that include behavioral, physiological, geospatial, temporal traces, and other biometrics in an adaptive Bayesnet. This is a very complex artificial intelligence system where the different models of behavior talk to each other. This forms a much, much better resolution identity for someone.

And it also predicts what they're going to do around things like credit. What are they going to do with their money? Or fraud, are they going to do something bad?

Since starting the company in January 2017, we've come a long way. We went through a pretty extended market discovery phase and still were able to sign our first paying customer within nine months of launch. We've been building out the technology analytics in our team.

Over the summer of 2018, we signed two more enterprise customers and now are poised to sign four more between November and the end of 2018. We're setting up for a tremendous 2019 with over 200 customer prospects in our pipeline and a lot of excitement around our predictive identity software technology.

The future of Distilled is quite bright. We are starting to see network customers come on. We are starting to see network customers come online. These are organizations that in turn have thousands of enterprise clients that seek putting our technology at the heart of their next generation offering and then working with us to secure scaling of revenue and customer footprint.

At the moment today, we are in conversations with network partners that touch over 2 billion consumers globally. We see an opportunity for Distilled Analytics to become the global standard for identity, credit, and fraud analytics. This is exciting for us because the whole point of this exercise is to generate change at scale, to solve structural problems in the financial system. And we're getting the partners online and the customers online who are going to help us do that.

[MUSIC PLAYING]

DAVID SHRIER: The partners we work with best tend to have a shared vision and a clear understanding of problem and use case. So for example, in looking at the false decline problem, we've had conversations with the leading credit card networks in the world. And they, like we, think it's a major problem. It's important for driving payments inclusion. It's important for helping prosperity in emerging economies, as well as helping the economy in developed markets.

And they believe that our technology can help crack the problem. They think that we can take false declines from 12% to 0% in developed markets and from as much as 73% to 0% in the emerging markets. That's a great partner for us, and we're really excited to pursue those conversations.

Other partners recognize that they have expertise in things like blockchain or payments technology in a particular market or region, and they need better biometrics and better analytics. They need predictive capabilities that can provide for credit models that don't need three years of credit history, but instead need three weeks of credit information or other behavioral information in order to build a prediction of what someone's going to do with their money.

So people who are open to embracing new ideas, who are ready to experiment and take technology forward, people who are interested in leapfrogging and going from the past to the future without stopping in the present, those are the best partners for us. So that includes everyone from a billion dollar credit union that we're working with currently on helping people promote better, more positive financial behaviors, to $100 billion insurance company that's looking to us to help them solve a major fraud problem.

Commercializing university technology is non-trivial. Even at the best universities in the world, and I hold him MIT up as the number one research university in the world, there's still a struggle to take that great scientific invention and turn it into a commercial tech. So that's been a journey that I've been following for several years. And I think we've found a pretty good methodology for determining product market fit and for adapting that university research into something that becomes a commercial product.

We use a discipline process. We talk to a lot of customers. And we go through an iteration again, and again, and again as we refine what we're hearing from the market with what the technology is capable of doing.

The other important thing that I think differentiates us from others is we continuously invest in innovation and in our university relations. So it's not enough to just find that one idea and then spin it out. You have to continuously invest in research to produce the best results and to keep ahead of the competition.

One of the things that we like doing is helping our large customers to innovate, to discover new revenue streams, new ways to cut costs, and new ways to get a leg up on the competition. Corporate innovation is not some mysterious process. I personally have worked with 11 different Fortune 1000 companies in helping them create new businesses and new products.

It can be systematized. It can be processized. It can be turned into something that's repeatable and scalable. And I think it's absolutely important.

And we look at some of the greatest successes coming from when large organizations partner with small, nimble startups that act as a bridge to whatever that future state is that they're trying to get to. I'm delighted that Distilled Analytics now has that opportunity to serve the financial services industry around bringing the customer of the future into the revenue stream of today.

[MUSIC PLAYING]

DAVID SHRIER: ILP is a fantastic resource for the Global 1000 companies that are trying to find that great innovation to the future. And speaking as a CEO of a small company, Distilled Analytics, I'm really happy that I have partners in the dialogue to navigate those large, complex organizations. Prior to starting Distilled, I was a lab group manager working with a professor at MIT and found ILP to be a tremendous collaborator in trying to bring some of the best ideas that we were coming up with out into the hands and into the minds of the various member companies.

Now that I'm part of STEX25, as a CEO, I'm thrilled that I have ILP's expertise and longstanding set of relationships helping me out in trying to find who wants to grow revenue and who wants to experience the benefits of these great technologies that we're commercializing. We're honored that Distilled Analytics was chosen as a STEX25 company by ILP. We think that we are industry ready. To us, that means we've made a significant investment in business processes and procedures.

In cybersecurity, we've invested hundreds of thousands of dollars in what's known as Soc 2 Readiness, which means that we have a third party audit company, BDO, that is certifying that our software and systems are built to enterprise class standards, and that we have a highly cyber secure environment in which we work with our data and analytics. But industry ready goes beyond just that. We have to be responsive to very large organizations and also understand how to work at their cadence. And that, I think, is something that we've successfully architected our business to.

We've also managed to bring in great collaborators like Ernst & Young to handle our audit, WilmerHale to handle our legal. We've really gone for best in class so that we are prepared to handle the needs of even the very largest companies on the planet.

Yeah, I can certainly talk about target clients. So the kinds of financial institutions that we're looking to collaborate with at this stage typically have between 5 billion and 50 billion of AUM, assets under management. That would be for a bank or a non-bank finance company who's looking to augment their capabilities with our software and services.

Other target clients that we're talking to include governments that are looking to incorporate our analytics into their electronic identity, or EID systems, and insurance companies, where the AUMs that we're looking at are actually a bit bigger. We are currently in final contract negotiations with a $100 billion insurance company where they see us as an integral partner to their analytics group in addressing issues like fraud and other forms of insurance activity.

Over the next couple of years, we are looking to bring our technology from POC into commercial adoption. We'll be doubling and tripling our team. We've been very successful in recruiting top flight engineering and analytics talent, and anticipate continuing to do so as we grow in scale. 5, 10 years from now, we'd like to see Distilled Analytics used by a billion people around the world.

We want to be the operating system of identity, credit fraud, and financial services globally. And that for us is what success would look like. Can we help a billion people get identity? Can we help a billion people get credit?

Can we help a billion people into the financial system? Can we help financial services organizations better serve those customers, and grow revenue, and improve profitability at the same time? That's our vision of success and where we'd like to see the future for Distilled Analytics.

Distilled Analytics is built around technology that was originally constructed to help the Global South, to help the poorest 3 and 1/2 billion people on the planet. One of the things we discovered along the way is that if you take those technologies that are engineered to help a billion people with an average balance of $10, and you bring them back to the developed world, to the G7 countries, you introduce a disruption in price and performance. That allows for business transformation in even some of the largest and most profitable markets today. Yeah, so--

SPEAKER 2: This is for my own interest. I'm curious.

DAVID SHRIER: So we have customers in countries like Brazil, or Mexico, or Colombia, or India, as well as customers in the US and Canada. The technology is incredibly powerful, and it's solving some very real problems for the financial services industry. So the companies that are the most excited about our technologies and our solution sets in our software platform are the companies that are global in scale.

[MUSIC PLAYING]

I'm David Shrier, the CEO of Distilled Analytics, a predictive identity software company that uses AI-driven-- shoot, I already messed it up. Try that again. I'm David Shrier, the founder and CEO of Distilled Analytics, a predictive identity software company that uses AI-driven block chain-enabled technology to solve critical problems in financial services, such as the $330 billion false decline problem.

It turns out that people using credit cards worldwide struggle to actually successfully engage in commerce. In the US, you may have a 99.0% approval rate when you try and use your credit card at a store. But in someplace like South Africa, for example, you might only have a 27% approval rate. In Indonesia, it might only be 45%. Globally, this lost commerce of legitimate customers who are trying to buy goods is over $330 billion a year.

One of the world's largest credit card companies thinks that our solution, Distilled Analytics predictive identity software, can improve the approval rate of these credit card transactions globally to that same 99.9% level we have in the US, unlocking $330 billion a year of commerce.

My co-founders include Professor Alex Pentland from MIT, one of the world's leading big data scientists, and Professor Alex Lipton, who wrote the Oxford Handbook of Credit Derivatives and previously was the chief quant at Bank of America Merrill Lynch.

We serve enterprise-class customers such as banks, credit unions, non-bank finance companies, as well as insurance companies and governments worldwide. We raised over $5 million of seed funding in 2017, and we are now raising a $10 million Series A financing.

[MUSIC PLAYING]