Vouched co-founder John Baird discusses ID Verification in the Age of AI: Deep Fakes and Reputation Management

SPEAKERS

John Baird, Rich Alterman

Rich Alterman  00:04

You're syncing up and tuning in to The Lending Link Podcast, powered by GDS Link. With a modern day lender can dive deeper into the future of Data Decisioning and Credit Risk Solutions. Welcome to the show everyone. I'm your host Rich Alterman, and today we're syncing up with John Baird, Co Founder and CEO of Vouched, a Seattle based ID verification company whose mission is to vouch for every important interaction on the internet by creating the largest identity and reputation management platform in the world. John's background and journey to Co Founder of Vouched is an interesting one and one we will learn more about in a few minutes. John is a graduate of the University of Missouri Columbia with a major in English Literature and Creative Writing. He is also an advisory board member of Wooter, a multimillion dollar direct to consumer sports apparel company. In this episode, John and I will touch on how ID verification solutions work, how they're leveraged in the financial service industries, the benefits AI and machine learning brings to these solutions and the challenges Chat GPT might present and so much more. But before we dive into the interview, please head over to our LinkedIn and Twitter pages at GDS Link, that's G D S L I N K and hit those like and follow buttons. If you haven't done so already, please subscribe to our podcast on Apple podcast, Spotify, or wherever you prefer to listen to your podcast. All right now let's get synced with GDS Link. Welcome, John. Hope you're having a great week.

John Baird  01:36
Hey, rich, it's great to be here.

Rich Alterman  01:38
Yeah, and thanks for joining me today. Where are you joining us from today?

John Baird  01:41
I'm actually joining you from a very sunny Victoria, Canada, up in British Columbia.

Rich Alterman  01:48
And what takes you to Colombia today.

John Baird  01:50
Being with some of our customers up here. So we've got quite a few customers across Canada, we're based in Seattle, we primarily serve customers, banks, financial services in the United States. But we've got quite a few customers in Canada and they are kind of peppered around the world as well.

Rich Alterman  02:06
Right. And I know we'll touch on that a little later. And thanks for that. So John, you co founded vouched in November 2019. Maybe you could spend a few minutes giving our audience a little bit on your background before starting bounce.

John Baird  02:17
Yeah, sure. So maybe a couple of personal anecdotes, it's a good place to start. I grew up in a small town in southern Missouri, it's called West Plains, which I always like to joke is neither on the plains nor west of them. But I could never figure out how our town got its name. Also home with scissors. That was the high school mascot, which apparently is an old name for a lightning bolt that was popular around the turn of the last century. So a little bit of Olivia, there for you. But I think it was in a lot of ways. It's what has shaped my viewpoint for Vouched in summers, I grew up hauling hay, that was how I made money winter dye was spent cutting wood. That's how we heated the house. And one of the things that I think about is just we have this real opportunity to enable people to be seen. And oftentimes, I think if you grew up in a small town, you feel like all the good things, all the cool things happen in the big towns. One of the things I love about technology is that it evens the playing field, right? There should be no difference between the ability and the access, whether you live in a small town in West Plains, Missouri, or you know, you live on the you live in the Bay Area. So I love that about technology.

Rich Alterman  03:29
So when you think about yourself as a human, how would you say growing up in a small town kind of influenced your personality?

John Baird  03:36
You know, I think if there's this great balance in a small town, because in a lot of ways you know, everyone, and maybe that cliche is a little bit overdone. Right you start in the town of my town was 8000 people, you certainly don't know, all 8000 people. But I think what is awesome is that because of the proximity, there's tremendous exposure to different people in different sort of walks of life, we definitely don't have the diversity that you would find in like a New York range, which I think is where you grew up memory serves. But there is an interaction with other individuals exposure that lets you understand I believe at a deeper level, different ways that you can take your life. And I also think that there's almost this unexpected pressure, good pressure, I happen to love pressure at the gate. You know, pressure produces diamonds pressure produces progress, to when you're seeing opportunity and that opportunity you believe growing in small town is often elsewhere. What it actually makes you understand is there's other things you can do and you can bring the best of that small town exposure and the learnings from that. And then take the learnings from other areas other opportunities and really bring those two together. And you know, sometimes you know some of our greatest leaders nothing against them. They often come from largest cities and they don't have, they don't have that exposure. So it gives me a very different perspective on things that I think is super powerful, and quite frankly, is one of the reasons why Vouched exists today.

Rich Alterman  05:13
It's interesting that you grew up in such a small town where maybe it was easy to actually vouch for everybody, you rely on a product where it's only designed to help in in a bigger echo system, where the person really is who they are.

John Baird  05:26
100% Can't really emphasize this enough, I think we'll get to this more later. But at the heart of every interaction is trust and safety. Right? In a small town to your point, you have a greater understanding of that person standing across the desk from you, right, like, both my parents were teachers, my dad was later a principal, I feel like they knew everybody in the entire town, they would see somebody who was 30 or 40. And they'd say, Oh, hey, Joe, or whomever I saw you at school when you were a kid, right? Or I knew your mom or I knew your kids, kids. And so there is this great community. And I think the real opportunity in a fully digital world is to really create trust at the center of all these interactions and transactions. And so, you know, I think that there's a vision out there where if we build the internet in the right way, the world actually starts to function, quite frankly, like a small town.

Rich Alterman  06:27
Right, right. It's interesting. I grew up in a town where we had a pretty big, small high school, I think the largest graduating class was like 154. And I have an older brother, Steve. And I think one of the downsides sometimes is you live in the shadow where I start a new class and the teacher say, Oh, I had your brother Steve. Alright, so expectation, he was actually a little smarter than me. The irony is that he makes his living with his voice. And here I am trying to catch up to him. So anyway, thanks for sharing some of those personal insights. Let's dive now into the world of ID verification. In my introduction, I shared the mission statement found on your website. And I want to remind the audience exactly what it said, because I think it's a pretty powerful one there. The company's mission is to vouch for every important interaction on the internet by creating the largest identity and reputation management platform in the world. You read that? And certainly you have set a very high bar for your company to achieve. Please share what was the catalyst for starting Vouched. And you kind of talked a little bit about it, but we'll dive into a little more detail, and any personal experiences that had you develop such a laudable mission? And once again, you touched on it a little bit, but let's dig a little deeper.

John Baird  07:31
Yeah, sure. I mean, society, civilization, right, its functions with trust at the center of it. And in the pre digital world, almost all trust based decisions were happening in person, they're happening across a desk. And oftentimes, in the old baking world, there's a lot of relationship that went often precede the compliance paperwork, if you will, digital rakes, all of that, and brings it pretty dramatically, simply because we are now making all of these decisions without necessarily knowing who is the person on the other side of that computer. Now, the great thing about digital is that we talked about this earlier, right? You no longer have to live in the biggest cities with the best hospitals are the biggest banks to opportunity, right, we can literally be sitting in West Plains, Missouri, and you can have a video chat with arguably the best doctors in the world, right? Or I can work with a bank, anywhere in the country, or potentially anywhere in the world. And then the challenge then becomes who is that individual? And so access is all about verifying who that person is. And so how can you do that, in a way that reflexively provides trust, both to the business, they make a financial organization, and that user that they feel trusted and valued as well, sitting on the other side of the table. And you want to be able to do that seamlessly, and securely. And what comes out of that, I believe, is economic empowerment, economic empowerment for people, no matter what time of day, it might be, no matter where they might live in the world. And I just think that that is spectacularly powerful. And I would hope for most people extremely encouraging. You know, one of the things that I think about, I think about this a lot, let's just say you're a single mom out there, right? And maybe you're working two jobs because you want the best for your kids. So you're working really, really hard, well, and you're working, let's say eight to five, and oh, by the way, you gotta take the kid to school and maybe a soccer practice after school, then you're gonna go get that second job. Why don't you go to the bank to open a bank account, right? Like, you don't have time. Banks are open 10 to five or whatever. They're open to maybe half a day on Saturday. And I can tell you, I could never make it to a bank during Baker's hours, right? And so my time to access the bank. And I figure if you're a single mom is Sunday morning, do a pm after you've been able to do all the other chores and errands and work that you need to do. And so Vouched is really here to help facilitate that in such a way that it provides access. And I would even say, equal access for people, no matter what their individual geography or constraints might be.

Rich Alterman  10:26
Yeah, that makes a lot of sense. One of the things that struck me when I read the mission statement, John, it said, the largest identity and reputation management, right, and I think about the ID side of your business, and that clearly aligns with identification. But when I think about the word reputation, if I was to poll most people and say, when I say reputation, what do you think about? And I think they would say, well, some measure, maybe not as articulate as this but a measure of how much respect admiration or trust a person or entity has earned, in particular social group or community talked about that, where you grew up? How do we think about reputation when we think about your offering, especially in the in the industries that we serve at GDS Link in the financial services? When people think about reputation, they probably think about credit? Does this person pay their bills on time? So how does reputation come into play here?

John Baird  11:12
Yeah, I think about reputation really, in two ways, there's a more literal way that is available today, right? As we talk about this, and then I think, where we are going, how I see the future of identity, and the future of identity is really the future trust, right? Worldwide trust because, again, it's trust is how we make these decisions. And one of the things that I'm very fond of saying I think it's absolutely true is your identity is your most valuable asset, because all value based decisions, or risk based decisions begin with verifying the identity of that individual. So to come back to your question, right on the reputation side, part of our evaluation of this person goes beyond just verifying their identity we have on the reputation side, we have an understanding or signals around risk, as well, because you might be who you say you are, right. But we can also tell that you have potentially been involved in certain questionable or even malicious activity. And these are some of the signals that we get back to that interview. But there's more this future state where I think we are going to get super excited about which is building a global identity network. And this is really the opportunity for the user not only to control their identity, but the ability to without them jumping through a whole bunch of hoops over and over and over again to get verified to be identified a single time. And the next time that they need to be re verified, it becomes a very simple process, again, with them in control of that process. This opens doors, right, it opens economic doors, it helps businesses to be able to do business so much more efficiently, and it injects trust into that echo system.

Rich Alterman  13:03
Now that makes sense. And if we think about even traveling, right, you start out with TSA PreCheck. And now you have Clear and digital ID. So the airports, I can show my face and makes things a lot quicker and more efficient. So I mentioned the financial services a few minutes ago, and probably obvious in that GDS Link clients are primarily in the financial services sector today, kind of talk, John about, you know, as you're selling into the industry, what are really those use cases and problems that you're helping lenders solve? You know, once again, I think it's pretty evident, but just want to give that opportunity to kind of level set.

John Baird  13:36
We really get ejected, generally speaking into four different use cases. Number one, if it's a bank account opens, right, we help digitally onboard and verify KYC the identity of that individuals who are part of that process there, you talked about lending, the number one way that we are used within lending? Is the verification of that person, especially if that is a fully remote loan that has been given to that individual but beyond just the credit, right? Does this person have good credit, we're really looking at is this person who they claim to be with a very high level of accuracy. And we can talk more about that the other two use cases that we see within financial services Vouched gets used in wire transfers. So the traditional method of calling an individual to see if they actually authorize that wire, I would say it's obsolete. Nobody should be doing that in 2023. It's extremely risky, even if you think you know, the person. And then the fourth way that we get used, it's actually in the call center, where somebody's calling in to they have a question about their account, or they need to reset the password on the app or something like that. How does that call center agent know who's on the other end of the phone? And so call centers are utilizing Vouched to ensure that the person who's calling in is who they claim to be and should have access to that account.

Rich Alterman  14:58
So John, you mentioned You're up in Canada and reworked in other countries. I think on your website, I read that you guys support over 1000 different identity types in over 150 countries, can you perhaps kind of like put them into logical categories or buckets of IDs. And in your experience and your team's experience, have there been any IDs that you've seen that you really felt were very unique and might be worth just sharing with the audience?

John Baird  15:22
Yeah, so we do work all over the world reaching that number is actually probably number of IDs, probably well, about 1500. At this point, might even be higher than that, I probably need to double check that number. At our core, we are an artificial intelligence machine learning company. And so when it comes to learning identity documents, we deploy a part of technology and AI a one shot learning and core to our proprietary AI. What that means is Vouched can learn look at a single document that is the source of truth, we know it's a good document a clear representation of the standards required by the government entity, or that identity document. And Vouched can study it and learn it very rapidly, usually within about 15 to 20 minutes with no human oversight on that. And so it's very easy for us to launch entirely into a new country very rapidly, sometimes within just a few days. So we can effectively master any identity document all over the world. And we do that, generally speaking, what we're seeing is a reversion, not a reversion. But most identity documents around the world start to look a lot like the US driver's license standard. Generally speaking, they're about the same size, they have a lot of the same watermarks, things that we would expect, they're all different. But there's a commonality or a general commonality to most of them. I will tell you, though, one of the identity documents that I have the most respect for is the Indian Aadhaar card, which is their national identity document, I believe that was originally rolled out about seven years ago, might might have my dates a little bit wrong there. But it is great as a source of truth and authentication of that individual, because it's very easily read. It includes biometric information on that person. So you really create the circle of authenticity, where you can check multiple aspects of that individual, including going back to the governmental source, as well as inputs from the owner of that ID. And it works for almost every single transaction across India. And so it's just a really great way to provide security to the business, as well as to that individual, so that they can transact in the way that they would expect.

Rich Alterman  17:36
So maybe it's getting a little unusual, since this is an audio podcast, and not a visual podcast, but maybe just if you could kind of give us a quick tour orally of the process that the applicant, let's say someone applying for a small personal loan, the process they go through with the VASH solution, I think that'd be interesting to kind of touch on.

John Baird  17:56
Yeah, sure. Let me try to do this briefly. So I don't get too deep in the weeds here, let's just say you know that an individual is applying for the loan, right, and they go to the bank's website, or the lenders website, they're gonna start an application process, they're on their laptop, and they're filling in whatever details that lender requires. And then at a point in the process, and this is going to vary according to the user flow of the company that we're working with. But at the right moment, that business is going to tell the individual, hey, as part of our security as part of regulatory controls, we need to verify your identity. And that's going to include looking at your face and looking at your ID, we will send a link to your phone. The user then gets a link to their phone via SMS. And they click on that link. And then on the web on the browser on the phone. This is not a native app, I want to be clear about that. Because we were very specific in not requiring a native app. Because no user wants to download an app just to complete, right some sort of onboarding or verification process. So they click that a page opens on their phone on the web. And with the user's permission, we then take over the camera Vouched takes over the camera, because Vouched can see and understand in real time. And then the following process takes place. So the first thing is we'll look at that user's face. And we deploy a liveness evaluation of that user. This involves the user doing a few simple movements, and these vary from customer to customer. And we evaluate those movements to make sure that they were done correctly. And so this is one of our security steps to help ensure that this user is real and alive. Then the next step is, as you mentioned earlier, Vouched is an expert on identity documents, driver's licenses, passports from all over the world. And so Vouched will visually evaluate the identity document of the user make sure that it has the criteria as required by the issuing government entity. And we will compare the photo of the user's face on that document back to the images of the users face that we took during the liveness test just a few seconds before, make sure that those match up. And then Vouched will read all of the information on that identity document, the driver's license number, their name, their address, date of birth, etc. And then we will take that information and compare it to public and private and proprietary databases, to see if it matches what is known about that user. And then we do a few dozen additional evaluations behind the scenes, other types of fraud and risk checks on that person takes me a while to walk through the whole thing. But it takes the user 30 ish seconds to go through seven, eight seconds to get the data back to the lender in this case, and that user is then approved and can complete that process.

Rich Alterman  20:59
Thanks for that. And I think it was a good walkthrough. So when you think about your clients, and we think about the type of loans done, and maybe the amount of loans, do you see where introducing that amount of friction? And it's not a lot of time? It's not a lot of effort. But still, it's friction, right? So do you find that there's a correlation between maybe the amount that a lender is giving out and the frequency that they want to dive that deeply with a product like yours? Or do you find it pretty ubiquitous, whether the lenders doing a $500 loan or a $50,000 Loan?

John Baird  21:31
Yeah, I mean, certainly, security and evaluation of an individual will vary based on the risk, right of whatever that transaction might be. I always say, if you're buying a Harry Potter book on Amazon, you don't need Vouched. You don't need us for data makes absolutely no sense. But I think for most lending decisions, Vouched does make sense now one of the things that we work with our customers very closely on is how and where does this evaluation happen? My background is actually in the consumer space. And so while I think about security every single day, I tend to lead with the user experience, because that user is your revenue, right? So I care very deeply that the user has a great experience. And the conversion of that user to completion is as high as is possible. And we're getting a completion rate of 97%. So in the high 90s, there, and then we're approving, on average, about 90% of those, some of those are fraudulent, some of those are being declined for various risk reasons. But completion rate of 97% is really what we aim to do. For everyone.

Rich Alterman  22:49
We talked a lot about lending, what other industries? Are you guys' having success in today?

John Baird  22:54
Yeah, it's really interesting to say that, you know, when I started the business, I was very focused on two areas, healthcare and financial services, because both have a regulatory requirement, verify the identity of the consumer right on the side of banking and financial services, and the patient will run the health care side. And really what I found, right, and we alluded to this earlier, you're creating online, we're creating a community of trust, for all of these transactions. And so what you now see are these numerous legacy industries going through digital transformation. And so we have large businesses now, certainly in automotive rentals, other vehicle rentals, RVs, for example, motorcycles, automotive sales, right? Every single time somebody sells a car online, they've got $40-50,000, going out the door, they want to make sure that that user is indeed who they they say they are. And we do quite a bit in real estate as well both with apartment rentals, verifying the identity of that person that is coming to the apartment to see the apartment. And there's risks, they're often directly for that apartment manager who's going to be showing that home, but also down the line, right? If that user needs to be approved for that apartment, part of the verification of the person's identity goes into their leasing rental application. And then we even do a little bit on the real estate side. And this is very nascent. But increasingly, I believe that there are opportunities where you would I could go view a home without the real estate agent being present, which I think is great, right? Like having gone through some of the boom real estate times in Seattle, where homes would disappear as soon as they were became available, the ability to be able to go see that home very rapidly. Could be critical if you actually want to purchase that home. But I think we'll see more and more of this kind of thing.

Rich Alterman  24:42
So have you approached any online dating services?

John Baird  24:46
It's funny, we've got a few that are getting ready to go live. Right now. Again, it all comes back to trust, right? Like all of these things that we would have done historically in person, and we would have had certain control was baked into those interactions, they wouldn't be as formalized, of course, as they are online. But we would have thought about things like you would know the person, generally speaking before you would go on a date with that individual, right, you probably would have interacted with them, maybe at work, maybe not at work, right, probably interacted with them before that first date. And so there's some level of trust that exists some evaluation of that person before you meet with that person, sort of one on one. And so certainly, I think dating is a big growth area. It's funny, I was just speaking to a very, very smart woman who covers the mental health and risks of teens. And we were actually talking about the risk of minors, on social media platforms, and their exposure to I'll just say, unethical adults, you can imagine, when we were kids, we all knew about Stranger danger, right? And we were at a playground, our moms would have never let us talk the Rando adult hanging out the periphery of the playground. And so you need certain similar controls in the social sphere. So your kid doesn't end up talking to the Rando that trying to interact with them online. And I think identity verification becomes key to the trust and safety of these types of platforms as well. The long and short of it is there almost an infinite number of use cases, for identity verification that we will be be implemented across the digital sphere, simply because that is how society operates. society operates on trust. And it needs to be a little more formalized in the digital world, simply because we don't have the the same processes in place that we had in the pre digital world.

Rich Alterman  26:50
Yeah, you got me thinking now about telemedicine because my parents, I've sat through a lot of tele med meetings with them. And we're trusting that that doctor is who he says he is on the other line of that camera, right. And I guess your ultimate success will be when up in the right corner. There's a Vouched UL stamp of approval, that the doctor really is who they say they are.

John Baird  27:11
To your point, yeah, it's a really, really big base. And we talked about how in the pre digital world, the physical world, right, you didn't necessarily have the kind of prescribed controls in place that you need in the digital world. But the reality is that there was a tremendous amount of formality and expense for us to understand. Is this a real doctor? You knew it was a real doctor? Because you would go see that doctor? At the hospital? Right? So the hospital itself was the stamp of approval, that this is a real doctor. And so now we just need to formalize that for the digital world.

Rich Alterman  27:52
So you mentioned earlier, your platform obviously takes advantage of AI and machine learning. But you also have the human element. Can you share with the audience today? What is that interaction between machine learning and AI, and your human staff, as it comes to, you know, improving, evaluating the models that you use to do the job that you do?

John Baird  28:14
Yeah, rich, if I could take a step back and just kind of talk you quickly through our approach to AI machine learning for Vouched. And then I want to bring it back to how we think about ongoing improvements, continuous learning for our models, when it comes to identity. There's more regulations all the time. There's, as we talked about earlier, right, like, numerous industries going through digital transformation, that need to understand who is that user on the other side, there's increased risk in the digital world. There's increased threats of broad, I think we're gonna talk more about those in just a minute. And I think what this all boils down to is increased identity complexity. If we went back 150 years, identity documents, generally speaking didn't really exist, right? How did you know who an individual was? Well, you know, that's Bob, I grew up with Bob, right, that was largely how it worked. Things were very, very relationship driven. You don't really see identity in the way that we define it today start to spring up until the early 1900s. This really gets accelerated with World War One, where you had a mass migration of people moving around the world, they were displaced refugees, because of World War One. And so you start to see countries say, Okay, we need some something in place to be able to manage this mass movement of individuals, right. And you're starting to see, you know, bigger rise of credit risk based decisions, right, that are being made at FICO. I believe it's created in the early 50s. If memory serves, I'm gonna have that date wrong. So apologies. But you start to see things like this right like more and more identity related information on an individual decreased creation of the social security number. Driver's license is existed I believe they were originally printed in the teens if memory serves, but the first one with a photo on it is not until 1956 in California. And why was that? Well, because a driver's license wasn't an identity document, it was a license to drive. And over time, we've sort of CO opted these other things, the social security number, the FICO score, your driver's license, we've taken all of this data to be able to stack them up and say, Oh, that's Rich, right? And we're cross correlating this information, were using multifactor ways of evaluating is this person indeed, who they claim to be. And there's more identity complexity today than there has ever been. And there will be a more into the future. This is not a single evaluation, a single data factor on a person that have a preamble to segue into what all this means when we think about how do we verify identity, there's a lot going on behind the scenes, with Vouched. So in an ideal world, it'd be pretty straightforward. We'd look at your face, we'd look at your ID, and we'd say, hey, that is good to go. But that's way too simplistic of an approach, right? The cardinality of fraud vectors is already massive today. And in addition to that, because of accelerating fraud, and increased regulation, there's an ever expanding visual and data factors that are required to evaluate identity. We deal with myriad variations of user inputs, identity, document format, document quality, we look at bogus evidence, device constraints, camera capability, even things like bandwidth limitation, as well as new types of emerging fraud. Deep fakes is a great example, just to name a few. So these are all the things that we're thinking about, for this user going through this process. And then from the evidence collection stage, to producing the identity verification result. It requires a sequence of multifactor AI, computer vision and machine learning models, as well as dozens of additional data factors. And we're running those synchronously while incrementally extracting and evaluating new signals related to that evaluation. And then we produce inference signals at each stage. And these are subsequently utilized in further stages and ensemble to produce the final result. That is, that's a pretty long explanation. And I haven't even gotten through half of what is going on behind the scenes from an artificial intelligence perspective to evaluate the identity of that individual.

Rich Alterman  32:48
I'll give you a plug. So that was a five minute gig session from John Baird with Vouched. Okay, let's kind of continue to keep an eye here on the time. I don't think any of us can get through a day today without hearing the word Chat GPT or Open AI we've read about a lot of potential benefits that will come from it, and perhaps an equal number of concerns that can come from its abuse. I recently viewed a video on an experiment performed by Joanna stern with the Wall Street Journal entitled, I clone myself with AI, she fooled my bank and my family. And I suggest everybody go watch it. While the clone had better successful in people and systems with a clone of her voice versus a video of her face, one can easily conclude that the technology will catch up. We're not even a system like Vouched, we'll be able to distinguish between the real and the deep fake. Am I wrong? I learned in classroom for you.

John Baird  33:38
Well, you definitely, well You know, I think the acceleration, accelerating pace of generative AI, the acceleration of fraud makes this problem really complicated. And Homeland Security. Actually, they have a paper out on AI and deep fakes in particular, and one of the things they say that every single financial institution should be thinking about is they say that deep fakes are so seductive, because humans can no longer trust their senses to verify the identity of an individual. Look, we've had 300,000 years of evolution, to that tells us that's trained us that what I see what I hear, is real. And now deep fakes have gotten so good. That what we see and what we hear, we have no idea. This is real. I was literally talking to my family about this over the past few days, that we should be very careful, even when we call one another about the information that we exchange. Because it may not actually be it may sound like me, but it may not be me on the phone. So only AI can actually manage this problem. Like a lot of the work that we do. It's no longer I mean, no longer just facial recognition, right like we've had that nail for years. The real IP around Vouched are all of the various fraud models, visual fraud models, etc, that we are using to be able to analyze, is this real? Or is this fake? And that's what everybody should be thinking about your traditional security. If you're still using things like Kba, just stop, I'm not a hacker, I can blow a hole through those. And seconds, you really need to be thinking about very sophisticated multivariate evaluations of that individual with AI at the core. And your AI partner needs to be evolving very, very rapidly, because it is a bit of an arms race.

Rich Alterman  35:39
No, it is, and hopefully, we can win it. So I'm going to take an opportunity I essentially like geeking out, I'll share with the audience that John had me go take a look at Google's Deep Mind. And as it relates to the game, go, if you're not familiar with a game go, it's a game that originally in China about 3000 years ago, and I kind of compare it to a combination chess checkers, if you're familiar with the game of fellow, and it has an astounding 10 to the power of 170 possible board configurations. I think that if I'm right, John, that goes out to certainly in the in the billions of potential moves. And what was really interesting about the article, and you mentioned the speed, is that in 2016, AlphaGo, which was their AI version, beat the top player, Lee Soto, the top world player, and their calculation was that they beat him a decade ahead of when they thought they could beat him. And now Alpha Go is actually the best player in the world. So John, anything you want to just add into that? Because I know you were kind enough to throw me on to that?

John Baird  36:43
Yeah, I would just say it speaks to the sophistication of AI, there's a stach, around the game of Go, that there are more potential moves than there are atoms in the universe. This is a complicated strategy game, right. And so the fact that almost a decade ago now, six, seven years ago, now, Google was able to create an AI to beat the world's best player. And go goes the sophistication of AI. And this is really before the current rise of what's referred to as generative AI certainly before the current news cycle around the newest versions of generative AI. And so AI is taking us into a very new place, technologically, and we're gonna see some amazing things, I believe, happened as part of this.

Rich Alterman  37:33
So one of the things that you hear anytime you're doing like either ID solutions, or really fraud mitigation solutions, they'll talk about concerns over false negatives, and as well as false positives, any lessons learned with your solutions? And how you kind of combat that issue of false positive and false negatives? And does your system allow the end user to kind of fine tune? How strict things are so that they can really eliminate the concern over false negatives or false positives?

John Baird  38:04
100%. But we talked about this all the time, right? Because the magic trick, the thing you're trying to pull off for your customers is, get 100% of the good guys through and keep 100% of the bad guys out. Right? False negatives, false positives. We talked about this literally, every single day. And it's the question on my mind, when I wake up in the morning, it is the thing that I'm thinking about before I go to sleep on our accuracy, we accurately verify the identity of 99.58% of the people that go through Vouched. And we've worked very diligently to get to this high level of accuracy on verifying those individuals to your other point, right like to maximize that user conversion and to be able to evaluate that person properly. According to the use case, I intentionally made Vouched extremely configurable identity verification is not a one size fits all, it should never be and quite frankly, you will, will never be done with this. So that's just configurable, both on the security settings and make it stricter, and make it a little less strict if you have less risk, right. And we have multiple ways of tuning to the security levels. We also have various workflows where we are doing a variety of potential step up evaluations on an individual. Again, these are matched over to the risk of the individual use case and the evaluation that needs to happen to that user, depending on where they are in that workflow and how initial evaluations of that user play out. We'll step them up or we won't step them up accordingly. And it's one of the big reasons why customers absolutely love us. We commonly have customers come to us and working with us, we will boost their conversion anywhere from 30% to as much as 3x versus the benchmark before Vouched.

Rich Alterman  40:01
So John, from a client perspective and the feedback loop, we often hear that lenders sometimes struggle to really determine whether a loss was a credit loss or fraud loss. Do your lenders contribute back a feedback loop to your system to help you further refine things by saying, Hey, we have these fraud losses.

John Baird  40:21
100% I love this, right? Like, the deeper that we work with our customers and deeper, we partner with them, everybody wins in that scenario. So what our customers and partners do is they're continually giving us data back, that this person was good, hey, we see something funky going on with this user. We take that information. And it goes into our continuous learning models, right. So we'll go with evaluate and study this feedback. And new artificial intelligence models come out of that our best fraud models have come out of this kind of feedback from our customers. And we always seek to get more and more data. And the other thing that comes out of it, quite frankly, is our customers are actively helping us to plan and roadmap, future versions future feature development, for Vouched. And it's a key reason why Vouched works in the way that it does.

Rich Alterman  41:18
Well, John, we're coming up on the time. And we're going to ended up having to leave lots of questions on the table. But this has been a really informative session. But before I close, and you kind of started touching on this, you mentioned that you were having some conversations with your family and saying, Well, maybe you can't necessarily trust that that's my voice on the other end. And, you know, we think about, we started out with a single authentication, then we went to dual factor authentication. And it's hard not to imagine that the number of things we need to do to make sure we're safe is going to just continue to increase. For our listeners, we're all consumers, we're all facing this problem, what additional words of guidance you give, given the industries that you support and the product that you offer? To help us all be smarter about how we are interacting on a day to day basis.

John Baird  42:03
So the biggest thing is just some level of common sense. And maybe slow down, just a little bit, especially when it comes to any sort of sharing of critical information, right, or security related information. I think, in some ways, we've all gotten a little bit complacent around texting somebody a credit card number, just for example, right? Like those are things where you want to take, take a beat, and you want to say hey, there's probably a better way to, to share this, when it comes to AI overall, there's gonna be some fraud, we all know that there's going to be some fraud, but at the same time, this is by no means the first and potentially not even the biggest technological revolution that we've gone through. I mean, we, you know, the PC revolution happened in the 80s, the Internet revolution in the 90s, we had the mobile revolution about a decade ago, we've gone through massive, massive changes, and we're all fine, right? We're all doing okay. I think humans are infinitely adaptable. We learn very, very quickly when we see new inputs or, you know, new potential threats. And we just scale up to be able to, to match those. So I think, you know, the biggest thing that people should do is, like I said, just take a take a breath, the big thing that fraudsters take advantage of, is time and stress, they make you feel like, you need to do this thing, and you need to do it really, really quickly. And they do that because your common sense goes out the window, there is no scenario out there where you need to immediately transfer all of your money out of your bank account into a different bank account, right? Like that does not exist. And I think people just know that they're going to be able to avoid 99% of the things that a fraudster throws at them.

Rich Alterman  43:50
And like you say, it's the same way lenders introduce a certain amount of friction, we can do that ourselves, right? And hey, you know, I'm not going to send you my full credit card number. I'll send you the first five digits, and then we gotta get on the phone together, something like that. But yeah, thanks, John. This has been really great. This is Rich Alterman, we've been syncing up with John Baird, the CEO and Co Founder of Vouched, thanks for tuning in and learning more about Vouched and ID verification solutions and trends. If you'd like to learn more about John's firm, please visit them at www.vouched.id That's vouched.id We hope you've enjoyed this podcast. Please stay connected with GDS Link and The Lending Link to listen to future podcasts and catch up on the ones who missed. Thank you and make it a great day. Thanks for listening. If you've enjoyed today's episode, please be sure to subscribe on Apple, Spotify, Google or wherever you listen to your podcasts. And be sure to leave us a review. Follow us on LinkedIn and connect with us on Twitter at GDS Link that's at G D S L I N K. Have a question for the show or have a specific topic you want us to cover. Hit the link in the description to drop us a note Thank you for lending us part of your day make it a great one.