Automating Document Processing and Building a Team in the Age of AI with Encaptures Will Robinson

Episode Overview

Will Robinson is the Chief Executive Officer of Encapture, a rapidly expanding SaaS platform that leverages machine learning to automate the extraction of critical information from documents for financial institutions. Under his leadership, Encapture has achieved significant milestones, including reaching $17.9 million in revenue in 2024 and serving over 50 customers, among them industry giants like Wells Fargo, Frost Bank, and Trist. These accomplishments underscore Will’s strategic vision and his commitment to driving innovation in the financial technology sector.

Before joining Encapture, Will honed his expertise in executive management roles at Dynata, a leading data and survey platform. His professional journey also includes experience as an investment professional at The Carlyle Group, one of the world’s largest and most successful investment firms. This diverse background has equipped Will with a comprehensive understanding of both the operational and financial aspects of technology-driven enterprises, enabling him to effectively steer Encapture’s growth and development.

Beyond his professional endeavors, Will is recognized for his thought leadership in the realms of artificial intelligence and automation within the financial sector. He frequently shares insights on building high-performing teams, fostering a culture of innovation, and navigating the complexities of AI integration in traditional banking environments. His forward-thinking approach and dedication to efficiency and customer-centric solutions continue to influence the industry’s evolution, positioning Encapture at the forefront of intelligent document processing. For a deeper insight into Will Robinson’s perspectives and leadership approach, you might find the following interview informative​

Episode Transcript

Will Robinson: The banks have to report a bunch of data about that loan. Was it accepted? Was it approved? Was it denied? What were the interest rates that were given? What are some of the demographic characteristics of the borrower? And so these compliance teams are tasked with reporting that data to the regulators so the regulators can come, you know, crunch the data and make sure everything’s good.

Kevin Rosenquist: Hey there. Welcome to Pay Pod, where we bring you conversations with the trailblazers shaping the future of payments and fintech. My name is Kevin Rosenquist. Thanks for listening. Will Robinson is CEO of Encapture, a SaaS platform designed to help banks and financial institutions automatically extract critical information from documents using machine learning. It simplifies document processing by capturing data efficiently and accurately, allowing banks to save time, reduce compliance risks, and improve overall operational efficiency. Will and I discuss automating document processing, leadership, Ship and building a successful team in the new age of AI. Joining me now from Dallas, Texas, Will Robinson. You’ve been in leadership roles for a good amount of your career. How much does you know? Company culture is such a buzzword and a buzz term and all that. Is that something that you feel is vital to leading a good team?

Will Robinson: Ye I think it’s critical, Kevin. You know, if you think about where the workplace is going, especially with some of the advancements in AI, you know, the the amount of people you have at your company or on your team is probably not going to grow if you especially if you’re in a growing organization, it’s probably not going to grow as much as it would have 5 or 10 years ago, because there’s so much more efficiency that can be that can be gained through automation. And so, you know, in my mind, kind of the only way you win is if you have the right people with a shared vision, being able to be highly effective, working together and performing at a high level. And so how is it that you get teams to be high performing? You have to have a, you know, quote unquote culture of in our company. It’s a culture of accountability, of trust, of high expectations. And if you can get people that can buy into that, you’ll have some you’ll have some people self-select out of that candidly. But if you can get people to buy into that, you end up raising the bar for everybody and you can get so much more done. And candidly, everybody has a lot more fun doing it.

Kevin Rosenquist: Yes, we’re obviously going to talk a lot about AI today. But you brought up, you know, AI as far as changing the way you build teams and changing the hiring process. Do you feel. Do you feel like that’s gone in a positive direction thus far, or is it easier or harder to build the teams with the AI being part of it?

Will Robinson: You have to have people who are willing to embrace it. And so the AI can be a tremendous resource for the people you have today, or the people you’re bringing on, if they’re willing to kind of leverage it and use it. And so what’s been fun at our company is that we have seen, you know, our sales team, for example, start leveraging ChatGPT to help do customer account research or to score deals to help better understand the health of our pipeline.  We’ve seen our development team start using AI as well around QA,  product development, code audits. And so what we’re seeing is, like, we have a great group of individuals who are good at what they do, but they realize they can be so much more productive and efficient and just get so much more done by leveraging this AI. So I think it’s only going to be good. I think, you know, if you have people who aren’t willing to embrace it, they’re probably in a few years not going to be as, as, as key as someone who can figure out how to be productive. And we even joke here, you know, sometimes, you know, when we were getting everybody signed up, it’s like, you know, 20 bucks a month a user for these different platforms. And my CTO was joking. He was like, you know what? If you were a smart employee here. You would just buy this yourself. Pay for it out of pocket. Use it yourself and just show how much more effective you can be. And then use that to get a promotion. Ask for more money, you know. Keep moving on your way versus wait for the company to come tell you to use it. This is something that everyone should leverage. So it’s a very big deal. I think it’s only going to get more helpful as we go forward.

Kevin Rosenquist: Well, hopefully all those people listen to this podcast and then they’ll all be running out to do that, huh? They’ll be running out to buy the product and get that promotion. Do you feel that, you know, in your experience thus far? Obviously it’s still fairly new for most of us. I and how that comes into the workplace. But do you see a lot of resistance to it from potential, you know, employees when you’re hiring, when you’re, you know, I mean, it feels like you kind of gotta be in on it if you’re going to be working in the tech world, especially with a company who actually focuses on AI quite a bit.

Will Robinson: Ye ye it’s funny because we, you know, so our product and we can talk about this in a bit, but but we’ve been leveraging AI for 5 or 6 years now. And it’s funny when we would go pitch our product to customers and prospects, mostly in the banking space, there were a few years where we would not even talk about it being AI, because it just had such a it would create such a visceral reaction from people that we talked about our software being more kind of automation, automation enabled and intelligent automation enabled versus AI words that could.

Kevin Rosenquist: Words they could handle, more words they can handle. Yeah. Well, as soon as.

Will Robinson: You say AI people freak out and they, you know, they say, well, you know, is this taking my job? Is this,  I don’t.

Kevin Rosenquist: Want robots being here and all that. Yeah. Yes totally.

Will Robinson: You know, or like, you know, especially within banking, they’re so focused on making sure that they’re not discriminating in their lending practices. They’re like, hey, if we introduce AI, is it going to make us biased in some way? And how does the AI, you know, make decisions and how do I track that. So some really good questions to be asked. But ultimately, you know, we look at the customers that have embraced our product as well as other AI products. And the amount of again, it goes back to just efficiency. And what that translates to is not necessarily job cutting. You know, we’re kind of operating in an economy right now where unemployment is really low. And finding really good people to work at your company is very, very, very hard. And so most organizations, they have a handful of really good people, and they want those people to stay and they want to, you know, create an opportunity for that person to grow. If they’re good at what they do, they want to grow. They want to take on more responsibility.   They also want to maybe eliminate some of the more mundane parts of the job so that that person can be excited about,  you know, whatever they do at the organization. And so I think I can solve a lot of those pain points in and around a great core team of people, and then you’re not having to look for that incremental hire, you know, every single time.

Will Robinson:  and so that’s where we’ve seen people use it. And I think, you know, for us it’s been, I mean, internally just running our own business. And I’ll talk about the development side, the product side for, you know, for example, we’ve had,  my CTO has been really good about this recently. You know, we’ll have an engineer go off and do something. It’ll take, you know, two days or so and he’ll come back, and then our CTO will pull up,  some, some sort of AI enabled product and show him how he did the job in 25 minutes. You know, he could have done it in 25 minutes, not two days. And so I think it’s kind of this,  a little bit of this is kind of a push, a push model where you show people the capability and you say, ye I guess you could do it yourself. But if I can be ten times faster doing it through AI and you can be ten times faster at some point, you either embrace it and use it and then leverage your own skill set that way, or we’ll find someone else who can.

Kevin Rosenquist: Ye ye it’s a good point. It’s a good point.So ye let’s go ahead and talk and capture.  You guys are positioned kind of at the intersection of AI and banking.  What do you see as the biggest misconception about AI’s role in the financial sector?

Will Robinson: There’s probably a couple things. One of them is probably true for every industry, which is what is this going to do for jobs? So our AI, we use it to process loan documents and so we can read incoming documents. Whether, you know, I give the example of applying for a mortgage because most people understand that. But when you go talk to a loan officer and you’re going to buy a house, they say, great, if you want a mortgage, we need to understand who you are, how much money you make, where your home is, and so you provide supporting documents to help them make that underwriting decision. That can be a pay stub,  a tax, a couple years of tax returns. You know, you need to provide a driver’s license so they know you’re a real person. And,  you know, typically there’s people in the back office that are just manually inputting that data and reviewing it to make sure they have all the data they need to go make a decision about whether or not to give you a mortgage. And I can do all that. And so, you know, there’s, there’s a lot of opportunity, I think, to use it in banking. I think some of the misconceptions people have is around, is this going to take away jobs? What we find most is most of our banks are understaffed in the first place, or they have people doing too many things and then focusing time on really mundane tasks like doing data entry or reviewing documents.

Will Robinson: And so, you know, if we can go automate that and provide a real lift, then they can take those people and have them do more kind of higher value activities, spend more time with customers, kind of solve the more, you know, work on, the more problematic loans, if you will, versus kind of the cut and dry loans that come through the bank. So I think that one misconception is that it’s automatically this big cost reduction play. And there’s probably a component to that. But really that’s in a growing organization, the cost takeout is that incremental cost that you don’t have to incur as, as you keep building your business. So that’s a big one. The second one is really around probably kind of bias and not wanting, you know, not understanding how AI works and saying, hey, look, we’re a bank and we have regulators who come in and they check every single step in our process to make sure that we’re being fair, to make sure that we’re not discriminating, that there’s no racial bias in the process. We’re giving everybody a fair shake. And that’s kind of a cornerstone to banking regulation. You know, how do we ensure that if we’re using AI, that we can show and demonstrate that we’re still being fair and we’re following rules and kind of giving,  giving our auditors,  comfort that, that we’re following all the rules and.

Will Robinson: Right. You know, my answer to that is it’s kind of the same way you do with a person. It’s like, you know, a person is making a decision based on a set of criteria. And typically, if a person makes a decision on whether or not to, to make a loan,  there’s supporting data around that that was used to make that decision. There’s usually like a, you know, someone else is double checking or there’s a committee approving this. It’s not just someone over in a corner making lending decisions. And so really, you can kind of treat AI the same way almost as this digital, you know, employee that is being fed data. It’s telling you. And it should be able to tell you how it came up with a decision that it made and that should still be checked by, you know, a person or committee kind of following normal, normal controls and processes. So that’s kind of a misconception that I think is a little bit overhyped, that somehow, you know, machines are going to make all these decisions for banks, and we’re not going to know how the machines are making decisions, and it could lead to a bunch of problematic outcomes. I just don’t think that’s going to happen.

Kevin Rosenquist: You brought up the jobs thing, too, and I think that’s interesting because I, I’ve, I’ve been in the marketing world most of my career. And here, you know, a lot of what the concern was when ChatGPT came out was replacing marketing people and of course, creatives being upset with how it was being used and understandably so, and all that stuff. But yet none of the job stuff, at least thus far, has really come true. And as long as you kind of have like to your point earlier about the employees and kind of embracing it, as long as you have an understanding of how to how to utilize it and make it, make it part of your, your kind of workflow, I think you’re pretty safe, at least thus far in most industries. Yeah.

Will Robinson: And it also And again, it kind of depends what kind of company you’re in. And if you’re just kind of a flat lined organization that’s not doing anything candidly, you’re not going to have great people working there in the first place. They’re not going to enjoy being there. And ye so I guess you could use AI to take out a bunch of costs, but you’re probably not even thinking that way because that’s just not the mindset that you have. Very true organization. Very true. If you’re a growing organization that’s trying to be innovative and trying to improve and holding yourself accountable to strong performance, you’re going to have these opportunities to introduce AI and get people around you. This goes back to the culture question, like you need to have a culture of people who are willing to work together and embrace this. And through that, what you’ll see is you’ll have you’ll have the team you have. And really, it’s almost like your top line opportunity is going to become that much greater. And so you’ll actually continue creating jobs and maybe at a faster clip than you otherwise would have without the AI. Ironically, even though it’s making you a lot more efficient. So net net, as this thing plays out, I do believe this is going to be a huge net jobs  creator. If you look at every meaningful technological advancement over the history of humankind, there’s always this,  a little bit of fear around it.

Kevin Rosenquist: Absolutely.

Will Robinson: Take away,  you know, jobs. And ultimately you look back and you realize it just unlocks so many more use cases across so many different industries that ultimately it leads to a lot more employment, a lot of growth.

Kevin Rosenquist: Very true. You can go back to all sorts of the different revolutions and see how it all, how it all actually helped more than hindered. You’re absolutely right about that. You brought up regulation a little bit ago.  you know, it’s definitely a concern with people,  striking that balance between, you know, innovation and compliance. How does Encapture strike that balance of driving innovation while understanding the compliance requirements that your clients have? You know.

Will Robinson: It’s funny, a lot of these banks have really, really good compliance teams that know the rules and the regulation. And instead of spending all their time using their big smart brains to go solve complex problems for the bank. They spend a lot of their day doing really mundane, kind of menial work that does have to get done. And so, for example, one use case for us,  there’s several different regulations to ensure that banks are not being discriminatory in their lending practices, and they’re practicing fair lending. One of the ways that regulators check this is they will ask the banks to submit all their loan data. So every loan or even loan application that came into the bank. The banks have to report a bunch of data about that loan. Was it accepted? Was it approved? Was it denied? What were the interest rates that were given? What are some of the demographic characteristics of the borrower? And so these compliance teams are tasked with reporting that data to the regulators so the regulators can come, you know, crunch the data and make sure everything’s good. Well, these compliance teams, they have to make sure that data is accurate because in an audit, if it’s not accurate, you can get fined. You have to go back and rereport all your data. It’s just a big egg on your face. So these compliance teams, very smart, talented people spend a lot of time just like looking through loan packages and double checking data, making sure that the data in their system matches, you know, the underlying source document or the PDF that the borrower signed.

Will Robinson: It’s really menial stuff. And so what’s cool about our solution is we can come in and automate that process. We can read through all the loan documents that were generated as part of the loan. We can find all that key data that the bank has to report. And then we can compare that data against maybe their system of record and flag any discrepancies and say, hey, 95% of the data you’re going to report, we found supporting evidence for in these documents. But in 5% of the data fields, there’s there’s two different answers. There’s different borrower names, there’s different addresses. Maybe it’s a different interest rate. And your document says 5% interest rate, but your loan system says 5.5%. Which one is it? You guys need to spend time figuring that out. And so we just cut down on a lot of that menial labor and kind of those tasks that are just not fun and trivial, but but still critically important for the bank. And the impact is those compliance folks now can really spend their time solving more complex, difficult problems, stuff that’s more fun for them anyways. And that actually has a strategic impact on the bank versus just doing the manual stare and compare.

Kevin Rosenquist: Well, there’s also of course the margin for error. I mean tired eyes looking at this stuff could be tired eyes on a Friday afternoon might not, you know, have might not be all that sharp. But the AI doesn’t have that problem.

Will Robinson: Ye the AI and that’s what’s funny too. You go back to misconceptions about AI. One of the things we get asked a lot is,  so the AI is probably 100% accurate. It can get everything 100%. And our answer is no. It can get 95, 97, 98%. But inevitably there’s going to be documents that it can’t find the data it’s looking for. Or maybe it can’t read your handwriting, or there’s just something about it and it can’t figure it out. And what you want the AI to do in that situation is kick it to a human to review it. What you don’t want the AI to do in that situation is guess and guess wrong, and then push the data through and say, oh great, we’re good to go. And so, you know, there’s been studies done on, you know, kind of false positive rates on people doing data entry. And we’re like way, way worse. You know, our false positive rate as humans is like 5 to 8%. So like 5 to 8% of all the data that we look at,  we type it in incorrectly. And that’s weird for us to think about. We think we’re, we’re 100% accurate, but we’re not good at it. You know, systems like ours, we’re sub 1% false positive rate. So, you know, we at the end of the day, we could be a lot more efficient. We can give you better data quality. And we’re not going to have to spend a little bit of time double checking the AI results if the system is not confident in its answers, but ultimately it’s still a much better outcome than you doing it yourself.

Kevin Rosenquist: Absolutely. Yeah. Looking ahead, what innovations do you foresee in document processing and automation? I mean, obviously automation is one of one of the top use cases for AI. And it’s just AI is obviously just going to keep advancing. Do you see anything on the horizon that’s got you fired up?

Will Robinson: I think it’s just going to keep getting easier. I mean, I think about even five years ago when we started really investing in AI, I think about the types of documents back then that we could process and process well, in terms of finding the data and extracting it. It’s become a lot more powerful. We can process a lot more what we call unstructured content. So documents that vary in their size and shape and layout format. Of course, a really rigid form that’s really easy. Anybody can do that. But if you’re reading, you know, a note that was written by a lawyer, those things are pages or paragraphs, and there’s all sorts of important information buried in there. But every note is going to be different. So can you train a system to go read through that and get you the data that you need? So that’s been really encouraging. I think it’s only going to become kind of more powerful. I think it’s going to become less expensive too. Over time. It’s just going to become cheaper. And, you know, really the big thing around it is adoption. Like, if you can really start getting enough organizations adopting this, you’re going to see really every like from a banking perspective, every borrower is going to start having a better experience with their financial institution because loans can get processed quicker, there’s less back and forth, there’s less waiting around. And that’s ultimately, I think, going to be better for, you know, for the borrower. And kind of the overall economy is we’re just going to be able to move a lot faster on decision making and do that with kind of better data.

Kevin Rosenquist: Ye that seems to be an emerging trend, an emerging trend that could really kind of redefine how financial institutions handle data is just the way that we’re, you know, the way that we’re implementing some of these AI products and to to make, like you said, the decision making faster. I mean, so much of what we talk about on this show and payments and fintech is just how AI and technology is going to make it so things aren’t they don’t take so damn long. It’s like so many things in the financial world take so long and like this is finally a real good sort of light at the end of the tunnel of Hope to to make this a lot faster. Yeah.

Will Robinson: No, I think that’s right. And it should percolate across every industry. And, you know, it’s hard in these regulated ones like, you know, you go to the hospital or the doctor’s office and you fill out 12 forms. And on every single form, it has your name, your Social Security number, your date of birth, you know, and you think like it’s got to be better. So I think there’s going to be a rate of change that is going to be slower in certain industries. But that’s the great thing about, you know, the economy and the world we live in is if you figure out a better way to do it, people will want to buy your product. And you can really go have a lot of success that way.

Kevin Rosenquist: You’re right. Like, we do seem like a society that is anxious or excited about even there. Even though there are a lot of people, the doomsday people with AI and all that stuff, you know, there are for the most part, we’re like, hey man, if it makes me more efficient, bring it on. I’m excited. Ye ye that’s right. And the larger sense, as AI becomes more prevalent in the financial sector, what ethical considerations do you think companies like, like even Encapture, need to be mindful of to maintain trust with their clients and just be safe overall? Yeah.

Will Robinson: I mean, I think being really clear on what you’re doing with the data that you’re being given. Banks are very nervous about and as they should be, you know, making sure that the data that they’re feeding into an AI system is being used responsibly, that it’s being protected from, you know, bad actors, it’s not being sold off or used against them competitively. So I think having transparency around, hey, if if, you know, Bank, you’re entrusting us to process really important information for you, sensitive information, how are we doing that and how are we safeguarding your data? That’s huge. So I think that’s one we talked a little bit about fair lending, just making sure that there’s no bias in the process. And some, some types of AI use cases kind of have to be a little bit more sensitive than, you know, the nice thing about our product? Like we are simply just pulling data out of documents. We’re not making decisions on behalf of the bank. We’re not, you know, trying to synthesize any of the data. We’re really kind of solving super basic stuff like, here’s a document, go find the name, go find the date of birth, go find the income, that sort of thing. And so,  I would say that kind of the flip side is maybe don’t paint every AI provider with the same brush. You know, candidly, we’ll get into a due diligence situation, a vendor due diligence at a bank. And as soon as they learn about our AI, like it, immediately flags it to this highest level of scrutiny. And, you know, and then kind of once we walk them through how our AI works and what we’re actually doing at the bank, they realize, okay, the risk here is very low.

Kevin Rosenquist: .

Will Robinson:  And you know, in capture is not going to be responsible for, you know, for creating bias in a lending process for example. So I think those are the two big in banking is like what are you doing with the data?  how are you protecting it? Are you doing with it only what the bank wants you to do with it? And then are you ensuring that there’s some sort of transparency in the decision making of the AI so that if a regulator were to come ask you, okay, hey, how did you get to this decision on this loan? And if I were involved, you had the ability to go explain it.

Kevin Rosenquist: Yeah. I’ve actually talked to people. I think it was kind of early on, but there’s people who feel like the AI is going to help with the bias, not hurt it. Would you would you land on that side too?

Will Robinson: It should. I mean, it should. Bias is really. Yes. This is to get real philosophical here. But you know, bias is usually based on the mind of the beholder. And, and if you’re if you’re using a system where there’s no discriminatory information provided, but, you know, into the system, you know, race, gender, whatever, and the system’s not, you know, sitting across from you across the table at a restaurant, you know, and making judgments about you. Right? That actually should eliminate a lot of bias that maybe we have as as human beings and even unintentional bias that a lot of us maybe are, you know, well-intended, but but still make decisions on, on like an appearance of somebody versus what are the what are the cold hard facts around your situation and what loan you’re getting and the fundamentals of either your business or your your personal financial situation. So I think there’s an opportunity there to kind of reset on that, you know, and as long as you’re showing that you’re not starting with like a bias system, in theory, as long as you’re not feeding it that information, you should be fine.

Kevin Rosenquist: Yeah. That makes a lot of sense. All right. Well Will Robinson within capture. Thanks so much for being here. Really appreciate your time.

Will Robinson: Yeah. Thanks so much Kevin I really enjoyed it.