Federal Tech Podcast with Brian Hettinger: Intelligent Automation for the Federal Government

In this episode, Brian Hettinger joined the Federal Tech Podcast to talk all about how federal governments can leverage innovative technology to improve customer experiences, prevent identity theft, and how identity affirmation isn't just about data—you need to have a multifaceted approach.

To get started with ABBYY Proof of Identity, please contact us here.

Episode transcript

This transcript has been edited for clarity.

John: Today we're gonna talk about identity management in the federal government. Welcome to the Federal Tech Podcast. My name is John Gilroy and I will be your moderator. Today, we have Brian Hettinger. He is a senior director for solutions, mobile at a company called ABBYY. Now, there are many companies I've interviewed in the past. And the reason I brought ABBYY on is because I was at a trade show and I heard some good news about ABBYY: they have innovation that can help my federal listeners.

And that's what this show is all about. I try to bring new topics on so people who are in the federal government trying to figure something out, they can have a new book on the shelf, they have another tool in the drawer they can pull out, and so the tool of yours today is Brian. Brian, I'm not calling you a tool. But that's (what) we're going to use (your knowledge) for; we're going to have you open up your world to the federal government so they can understand a little bit how ABBYY can help them.

But before we begin, perhaps you can give me just a thumbnail sketch of ABBYY and where they fit in this whole world of identity management.

Brian: Absolutely. Thank you, John. Really appreciate your time today and all your listeners' time today. ABBYY is actually a leader in intelligent document processing. We're a company that's been around for a long time. And we've expanded our role in intelligent document processing into identity proofing. This whole concept of document-centric identity proofing and identity affirmation is really hot right now; Gartner recommends that these two solutions be combined. And frankly, in the marketplace, there are very few vendors that are offering any kind of technology like this. So this is a key area for ABBYY, we're helping federal governments, state governments, and all other industries as well. It's an exciting area for us. Thank you, John.

John: Talk about excitement. I'm gonna talk about an exciting movie Back to the Future, because we're gonna go back to the (past then) future. Brian 20 years ago, this is hard for you to understand this, you would waltz into an agency, let's say interior, you get a little Personal Identity Verification (PIV) card. You know, (you) waltz right in and and then down the hall, there was a server. So, what's the problem? My problem here, things have slightly changed in the last five years.

So, now we have John Gilroy (and) I'm not waltzing into the office, I'm working remotely, and the servers (that were) down the hall, that server could be in Denver (or somehwere else). And so a lot of problems present themselves. That didn't happen 20 years ago - Back to the Future today, Brian.

Brian: That is so true, John, so many things have changed. You know, many of us, in fact, I can't even remember what life before the iPhone was and how it's something we're all used to. I even had the opportunity, John, in my past of bringing this whole concept of mobile check processing—being able to take a picture of your check from your phone rather than having to go into a branch—and I was one of the few individuals that brought that to market. And, you know, initially they thought we were crazy, you know, and now it's just common, it's something that everybody uses.

It's a similar kind of technology that we're doing today. But using it in the identity space, along with all kinds of other documents associated with identity to be able to prove you are who you say you are, remotely.

John: But it's just so crazy. I have a daughter living in rural Africa. Someone sent a check here to her. So I took a picture of it, sent it to her, and she deposited in her bank in Africa. It's like, "Whatever happened to walk into the bank, and it closes at two and bankers hours?" It's a different world isn't Brian?

Brian: It is it is a very different world we live in. We're in this world, where you have these big tech companies, Apple and Amazon, (that) have created these experiences. And people are not only getting used to it, they're demanding it. They really expect that and the federal government has to step up.

So the world has changed. But you know, there's a lot of opportunity within the federal government and government agencies to be able to adopt this kind of technology and make it a win win for everybody. Because it isn't just about the end user experience. I mean, certainly that's a key part. But it's also about improving the process for federal government employees that are having to deal with this on the other side of the aisle too, or in the brick and mortar. So yeah, certainly things have changed.

John: I live in a small town in Virginia, they have a barbecue place with a secret sauce. And so my question to you is what is ABBYY's secret sauce? For my federal listeners, what is a differentiator that you bring that other companies can't?

Brian: That's a great question, John. So one of the challenges today, or one of the biggest challenges today is, is really what I call kind of pieces and parts. Today, there is technology that exists, that technology has evolved, if we were having this conversation, maybe five or 10 years ago, the technology wouldn't have been mature enough. But the technology now exists today to be able to do all these different components. The challenge is that typically for a government entity to implement this identity proofing and document-centric identity affirmation technology, they have to work with maybe five different vendors. Well, how are you going to do that? How are you going to research all these different vendors? How are you going to figure out who's the best then on top of it? Once you've done that, how are you going to have ongoing work with them, update all their different systems? And how are you going to gain the knowledge to make them all work together seamlessly, and in a very efficient and and accurate manner?

Well, ABBYY has actually done all that hard work. We're one of the very few vendors out there that are offering a complete solution of document centric identity proofing and identity affirmation. So we're combining all these characteristics of identity proofing—you have ID reading, ID verification, facial matching, and liveness detection—combining that with all these other components like trailing document processing, being able to process any particular document, for your specific use case (maybe it's a birth certificate, or it's some other documents specific to that particular need or use case of this situation) and being able to do that. Having process intelligence and having things like dynamic fraud detection, bringing all of these technologies together, John, in a way that provides an easy solution that's easy to implement. This ultimately ends up saving a bunch of money and allowing end users to be able to go through this onboarding process, or go through this process remotely, rather than always having to go into those brick and mortar locations. 

John: So there's a concept you tossed out casually, I don't know how its trippingly on the tongue, maybe - it's called identity proofing. Can you define that for our audience?

Brian: Absolutely. Thank you, John. Identity proofing is really about trying to understand who someone is and making sure they are who they say they are. So identity proofing in itself, is typically multiple components.
Typically, you're needing to take some identity document, the simplest form would be a driver's license or your passport or something like that. Then being able to not only recognize it, but extract all of the data off of that document. Often the first step in the process is maybe pre-filling a form so we're able to use this smart auto-capture technology. All a user has to do is take their mobile phone and point it at their passport, for example. And it's automatically going to capture that image when all the conditions are right.
So they're not getting a blurry image—everything looks good. We extract all that data. Now we can pre-populate the form maybe that they're using for some onboarding process.

But then it doesn't stop there, you have to understand—is this truly an authentic document? There's something called ID verification. So is this really authentic? Or has it been artificially created or tampered with, or is it a fraud? And it doesn't stop there, because typically, the headshot from that ID document is extracted. And now we ask the users to do something most people are familiar with: take a selfie. And through that process, we're able to do facial matching—we use biometric technology to do facial matching, to ensure you are who you say you are. Because I could have your ID John, and it's a perfectly valid ID. But as soon as we do facial matching, that's not going to succeed because I'm not you. So being able to have all those different identity proofing components are really important (especially) when you're trying to do something remotely and not in an in person environment. But of course, identity proofing is only part of it, but it is a key part of the process.

John: Motivation is important for distance athletes, for the people that run the Boston Marathon, and I think motivation is important for changing technology. You know, if my daughter would live three doors down from me, I wouldn't have taken a picture of the check, I would've walked down the street, and handed to her. So let's go back to the beginning of the show here. You know, if we, you and I, worked in the same office, and I got thrown an email, I could walk down to your desk to you say, "here, take this". But I think what happened, COVID has forced people to to look at new solutions. And maybe it's forced people to look at ABBYY and go, "Well, you know, my daughter's in Africa now, or I've got to commute 60 miles, I'd rather commute remotely". And so maybe we have to look at this and see, maybe that's a way we can handle some of these documents. Is that what you've seen the last few years? 

Brian: That's exactly what we've seen, John, it's really interesting, because there are two sides to that—COVID really accelerated this whole remote process. One of the two sides is in terms of citizens. And, you know, we can't say offices are closed, we can't necessarily just go into some brick and mortar location. There were times when that simply wasn't possible. So how do we still meet the needs? Not to mention, generally, none of us want to go into some brick and mortar location stand in line, (and after waiting) they say, "Oh, you need x and y documents" and we're like, "Oh, well, that's a bummer. We didn't bring all those documents". And now we have to go home, (get the documents) and come back to stand in line again. And that's not generally a friendly process. So certainly, that's something to be avoided.

But the other side of this is for the employees. Think about the federal employees where their office was closed, well, how do they still need to do that work? And so through intelligent automation technology, we're able to take some of the work that they had to do in person, in the brick and mortar location, and automate it to where, now from the comfort of their own home through a secure connection, they can simply review and have a human-in-the-loop automation for some of these processes. Therefore, rather than having to be so engaged, they can actually be much more productive.

You know, we actually did a survey recently, John, and what was interesting about this survey is that in government applications, on average, one in five people drop out. And the main factors contributing to that is that about a third say, the biggest factor is the process takes too long. Another third say that identity proofing and affirmation is a key pain point. And the last third, there's simply too many steps. It is a complicated process.

Certainly COVID did accelerate all these things. But it also kind of showed these pain points and really "ripped the band-aid off" the situation because there was a lot of organizations that had plans to to improve the situation, but all of a sudden, you know, it hit and they really had to accelerate those or be faced with some of these challenges. ABBYY has been helping organizations address all these challenges and improve this situation, not only for that end user, that citizen experience, but also for their employees to be able to work remotely, more efficiently, and be able to handle more volumes and actually do more higher-value tasks, as well.

John: Half of this show, I had this silly image of me walking into the Department of Interior (DOI) and I got my PIV, everyone's happy. Well, what happens if you're at FEMA (Federal Emergency Management Agency), and there's a flood somewhere and you have to get federal assistance? Well, do you have a PIV card? I didn't plan on getting flooded. And so there's situations that I think are highlighted by COVID—but it's always been there. I mean, there'd been floods in United States for 300 years, 200 years, and so on. (It's) nothing new and so I think what could happen is, if the federal government says "you have to improve citizen services that have better user interface", well, then one way (to do that) has got to be a more efficient identity management. I mean, that seems like it's hand-in-hand with serving the citizens, doesn't it? 

Brian: That's exactly right. You know, one of the areas that there is a lot of fraud is actually with federal benefits—people that are getting benefits that actually shouldn't be getting benefits. And part of it was because of COVID. Maybe they weren't going into some brick and mortar location, maybe they didn't have the process in place to be able to check all those things. So, you know, what we're finding, John, is that there are certain use cases right now.

For example, one of the things in the US is this initiative called Real ID, where we can essentially get a new ID that is basically a combination of a driver's license and federal (identification) all in one. And what what we're finding is that, because there's some timing associated with getting your REAL ID and being able to have that federal document to be able to fly in the US and other areas, there's some initiatives to get people this, but they have to go into brick and mortar location. It's complicated. So we're working with governments right now to be able to do that remotely.

But what's interesting is, although that's the driver right now, simply because of the timing, many of these government agencies already have plans to expand the same process, the same onboarding process, the same identity proofing and identity affirmation, through using this document-centric approach that we've been talking about—(not just these), but for all these other processes, (too). It could be simple things like getting a fishing and game license, there are all kinds of other different things that local, state, or the federal government can offer. But certainly, there are areas that are more prone to abuse, or maybe have higher dollar values associated with them. Or maybe there's more urgency around them for certain claims. Those few areas are really key areas where this technology is frankly a no brainer. You know, it fits great, and it is a win win for everyone. 

John: Well, Brian, you have a degree in engineering, but you're very carefully phrased that sentence there. It was like a PhD in English. I mean, we can just look down at Florida and go fraud and abuse Florida. Okay. I mean, we can just look at the zip codes. So I'll say you can be more discreet about that. But I think we all know what's going on. And this isn't a couple of hundred dollars. This is billions of dollars. And people are scamming the government. It's just a terrible situation.

Brian: It really is John. And the other side of that, too, is identity theft. It's impacting millions, billions of people. There's all kinds of fraud. There's all kinds of fines because of fraud. I mean, it's just it's crazy the situation, not only in the US, but globally. And not only in the federal government, but in the private sector as well.

John: Well, you said the magic word federal government, and I'm in the Washington DC area, you're not. In Washington DC, we have a mandate every 13 minutes, we have to say zero trust. So if I'm buying apples, because the zero trust, I'm gonna change the courtesy. So I gotta say, zero trust. If you look at zero trust, which is a big push to the federal government now, here's the six pillars—here's the six pillars of zero trust the seven pillars, here's the five pillars, every pillar—the first one is identification. I mean, that's the that's the commonality. And there that could be six or seven, or I don't know how many pillars, have to study Greek architecture to figure out how many pillars there are. But it starts with identification.

So what you're saying is that it is possible that the solution from ABBYY might be able to help my listeners, get a handle on this whole process of identification as the first step to this huge amorphous picture of zero trust. Is that right?

Brian: That's right, John. The key thing is one of the things that you just said, it's really important, and that is: there isn't just one pillar, there are multiple things. That's what not only we have recognized, but even analysts like Gartner. It isn't just about doing ID verification, it isn't just about trying to do facial matching, it isn't just about these trailing documents, and being able to read other other things and having this identity affirmation—it isn't just about data. Frankly, it means most all of those things, you know, you need to have a multifaceted approach.

There's an example, John, that I like to use because for me it's quite personal and sort of hits home with me. I've been part of some processes, where they don't have a very good balance, I would say of technology and human interaction. What I mean by that they'll ask me to fill in some information on a form. And then rather than asking for more information, they'll start asking me some questions. And these questions are things like, "what was the color of my car 15 years ago"? Well, you know, maybe I'm exaggerating a little bit, but you get the idea. It's a little bit weird that (this system is) asking me that. And so one of the things we really recommend is that you engage the citizens they have in it.

They have their birth certificate, they have these documents, and that will engage them in the process, even though it's required or it's best practice to do multiple things, it doesn't really matter if I need to take a picture of one, two, or three different documents. I can still do that in seconds, or you know, just a minute or two. And yet I feel engaged in the process. Taking a selfie is something that most people are used to doing. And so it's a natural, easy part. And yet, you're able to knock off lots of different steps and lots of different things to ensure the process is very secure. You're also doing a lot of things to eliminate fraud in and, you know, if you want, we can chat about it more. But ultimately, this is something that's good for everybody, because it also reduced reduces the cost of of transactions. So this isn't something that that should be seen as difficult. It's great for end users, and it's great for employees and the government.

John: I've interviewed a lot of different people. Couple months back to I interviewed a PhD in math, Dr. Elsa Schroeder, and she is on episode 40—if you want to listen it's called the rise of machines and federal technology, we talked about mathematics and machine learning, and a lot of pretty interesting topics. Which brings me to identification. That brings me to the robots. I mean, there are non-personalities out there that need identification as well, don't they? I mean, this is almost like science fiction. I know, but it's true. I mean, there are systems that have NPE's (National Electric Mobility Platform), that need to be identified. So where do you fit in with a federal problem inherent with processing robotic process automation, automation identities that are non-human? It seems like a science fiction question. But it's a real question, isn't it?

Brian: It is there's all kinds of things and certainly the robotic process automation is a key part. But typically, when you look at what, what we're doing is we're really in the first part of of the process. So oftentimes, the key initial step is this onboarding process to determine who you are, and if you are a real user. And I need to not only understand that, but I need to affirm that. And that's really where our part plays the biggest part, John. So when we're onboarding humans into some some process, now, after, after that occurs, and certainly, if maybe there's ongoing services, there are other techniques and other technologies that you're wanting, you're going to want to do to be able to affirm that in subsequent uses. But, frankly, fraud is sophisticated these days.

So, when you talk of automation, you talk of what robots can do there. You could even have a robot that has some fake ID document, and also has some fake picture of somebody. And it could be trying to use that or some fraudster could be trying to do that as well. The technology that we use is so advanced, John, that it's actually using liveness detection. Although I mentioned that it's doing something as simple as having someone take their selfie, okay, the idea is that it is very simple for the end user, but what's actually going on behind the scenes is very complex. With a selfie, what happens is this liveness detection is making sure this is not a picture, this isn't just some fake robot, or fraudster—instead this is actually a human that is alive and breathing. And real. Only if it passes that liveness detection, then does it actually go on to be able to do the facial matching components, and make sure truly you are who you say you are. So that is that is a big part of it. There's this technology and some of these things I mentioned sound simple, but in fact, the technology behind it is quite complex. Whether it be facial matching technology, NIST (National Institute of Standards and Technology) standards, and all of these kinds of certifications, etc. But even the reading the documents, because it's not just about extracting a little bit of data. It's it's about truly understanding what is this document. What is the data that I'm extracting? Is this a name or a date, or this is a birthdate not an expiration date? It's truly understanding what all this information, all this data is, and being able to leverage and use that to make sure this isn't some fraudulent, some bot, etc. But no, this is instead some real life individual doing some proper process.

John: Brian, time to use two four-letter words, are you ready, FCC (Federal Communications Commission) is gonna come down on us. And the words are deep fake. I mean, I imagine your technology can be used to recognize deep fakes. You know, I have a friend that says, he plays around with this all the time. It's hilarious. I think it's really funny, until I realized that people are good at it, aren't they? 

Brian: They are, it's exactly right, John. Deep fake is a big deal and the technology has also advanced. One of the things that I really like about ABBYY's technology is that it's easy. As I mentioned before, in the past to be able to stop this deep fake type of technology, you had to do a quite complex process; it wasn't just about taking a selfie, you had to basically take your selfie but then you had to rotate your head and maybe had to blink. It required this complex process, which, frankly, is a challenge and leads to abandonment. And not everybody is going to actually follow the instructions. So the technology has advanced enough to where now we can do simple things like take a selfie, or to something very easy to the end user, and you leverage this advanced technology to prevent deep fakes. Yeah, absolutely.

John: I think of intelligent document processing from years ago. And that meant scanning something. Now this has gone like from the Wright brothers to going to the moon, it's like Ray Kurzweil from a few years ago. And it was a pretty big deal. It's a nice trick. It's incredible the amount of sophistication you could have with this information. Now, it's almost at the cutting-edge. And there may be people that resist this, because it's just, it's unbelievable in some ways, isn't it? 

Brian: It is, but the thing that I would encourage users to understand and everybody that's listening to understand, is that if we were having this conversation five or 10 years ago, yeah, it would be to cutting edge. And frankly, what I like to call sometimes "bleeding edge". But it is advanced. And you know, now when we think about what happens in document processing, you're right, that artificial intelligence, the machine learning, it's also doing natural language processing. So it can understand and read information within the documents and actually know what that information is, know where to put that information, know what to do with it, just like a human would.

There's other areas too, like, like process intelligence. This is a key area for us as well in process mining, because what it what it can do is it can go in and look at the larger process and find bottlenecks that occur. As humans, you know, we don't even do this on purpose, but sometimes we have a bias or maybe an opinion or something. And, frankly, what we found in many cases is using process mining and process intelligence uncovers things that the the humans, the rest of us or the rest of the people close to it, didn't even know what was going on. So there's a way for in the efficiency there that can be uncovered. And then as we're talking about ABBYY Proof of Identity, there's this advanced artificial intelligence, whether it's looking at the ID and looking at all the security features or being able to look at little tiny things that even our human eyes can't even detect, that may have been manipulated in a document and using technology to do that. We talked about deep fakes and selfie/facial matching and all of that biometrics, so yeah, the technology has very much advanced—but to the to the citizen, you know, that's maybe looking for some benefits or maybe they're being on boarded for some other new process, it's a very simple experience for them. They're not exposed to layers of complexity going on.

But what's also cool with this technology, John, is we're able to use all that automation and technology and are able to do this within seconds. It isn't something that takes, you know, minutes or hours to process. Certainly if you have the old way, or the current way, where typically somebody has to go through that stack of papers—that's going to take hours or days or weeks. They're usually overworked. But the computer can scale up, the server's can scale, and they can do it in seconds. So yeah, two years

John: Two years ago I had a student working for Deloitte, and he handed me his business card—he had UI/UX on his business card. And I said, "Ooh, this is pretty trendy". And when I think about what you just said, if you take the last half hour and have a transcript of and highlight all the important things, it seems like, the technology is fine. And you know, intelligent automation is fine. But the end here is user experience. It's like, okay, this is a guy in the VA (United States Department of Veterans Affairs) can get his claim done. Someone can get a FEMA trailer, someone can get a loan for flood relief, it seems like if you keep your eye on the the end of the race, or then the races here, we can dazzle you with all this fancy talk and everything else. But if you can give a better user experience that's going to help competition, agencies goals. It seems like that's we're talking about, isn't it?

Brian: I agree, John, you know, we usually talk about a few different things, you know, some some might approach it from a fraud perspective, some might approach it from a cost reduction perspective. Certainly, there's some numbers and statistics out there where, you know, doing something in a brick and mortar is much more expensive when someone walks in to do that transaction versus doing it remotely. And that's even been proving back to the checks that we talked about early on. So there's a few different areas. But at the end of the day, usually the most compelling one, and usually the one that hits all of us at home, close to our hearts: we don't want to stand in line. We want to have a great user experience all these other technology companies, etc. have given us this instant home situation where we can do things very easily from our mobile phone sitting from the comfort of our home, and do what we need. And being able to do that in federal and government agencies is  key. And that is really what it's about having that great user experience.

John: Well, our users had a good experience listening to you. And so I want to thank you for your time here. It's a whole new world for ABBYY. Many people have heard about it—go to the website. They have articles about success they've had in many aspects of helping federal agencies reach their goals. You have been listening to the Federal Tech Podcast with John Gilroy. I'd like to thank my guest, Brian Hettinger, Senior Director of Solutions, Mobile at ABBYY.

Read our checklist for five steps to intelligent document processing.

Download your free checklist