Michael Purifoy, chief treasury and digital banking officer at VeraBank, joins Adam Blue at Q2's CONNECT 26 conference for a candid, on-the-floor conversation about where the technology is earning its keep and where the real work is still ahead. From an 18-month process mapping exercise that surfaced two years of workflow debt, to the harder challenge of helping employees see change as an opportunity rather than a threat, Michael shares what disciplined, customer-first AI deployment looks like in practice.
Adam Blue
Hey everyone. Welcome to Cut to Context. We are here at CONNECT. That's why I'm wearing this goofy badge that they make me wear so I can remember who I am. I'm joined by Michael Purifoy, chief treasury and banking officer for VeraBank. And we are going to get into AI where it really meets reality in his bank at VeraBank today. So really excited to be here with us today, Michael.
Michael Purifoy
Yeah.
Adam Blue
Why don't we just get right into it? What would you characterize as the position you are on the journey of using AI at VeraBank?
Michael Purifoy
It's a good question. I think at this point, we're selectively deploying technology. I mean, we've been trying to do our best to track as the technology has evolved. We've had our RPA in place since 2015. We feel like today it's doing the work of probably five FTE.
But I think here recently, as we've deployed LLMs and agentic technology, we've been partnering with some of our big vendors to put that in place. We have an LLM knowledge base that our employees are using today. We put that in production at the beginning of the year. And we're also working with Q2 right now, testing and about to roll out Q2 Assistant, and then some other vendors that we're working with, to try to push the technology forward.
Adam Blue
Well, that's fantastic.
Michael Purifoy
Yeah.
Adam Blue
Why don't you take me through—not Q2 Assistant, because we'll probably talk about that a little bit later—why don't you take me through a use case that you've built at the bank that's kind of homegrown that you're really excited about, some of the payback on?
Michael Purifoy
Yeah. So here recently, we worked with a third-party consultant to sit back instead of leading with the technology. We wanted to figure out where we had some problems. And I don't know if you've ever experienced this, but nobody likes calling their baby ugly.
Adam Blue
Oh, I've never had that problem.
Michael Purifoy
Yeah, yeah. And so 18 months ago, we started trying to map out some of these processes and we were having a hard time getting them across the finish line because there was some resistance. It's like, well, we don't want you to change this or that, whatever. So we partnered with that third party, and we knew going into it—our chief technology officer or myself—we were like, we are going to find some processes that are pretty hideous. And we did. And we looked across the entire bank. It wasn't just in one area. It was commercial, retail, consumer onboarding, audit, backroom deposit ops, all kinds of different areas, treasury, wealth. And at this point, what we're seeing is that we probably have two years' worth of work to try to build and plug this modern technology into our workflows.
One specific use case: It's funny because there's definitely some resistance to our employees accepting the change that this is going to bring. But they tend to complain about the repetitive work that they're doing. And so we had one employee that was responding to subpoena work. And literally when we said, "OK, show us the clicks, show us what you're doing." We realized like every bit of this work that she hates can be automated, and we can speed this up and move her to the work that she really likes doing. And that's working directly with customers and not spending days upon days responding to a subpoena. So that's one use case. I mean, we probably have 200 or more that we've identified just by looking within our workflows.
Adam Blue
So that's a fantastic catalog of your process debt in the organization. That's interesting. It's funny you bring that up. So you have an employee—I'm sure she's a great employee. She probably cares very much about customers and doing the job and her coworkers. And she tenaciously clings to the sinking life raft of this terrible thing that she does every day. And there's this—I don't know what it is—it's almost like some version of Stockholm Syndrome in the office where people fall very in love with the terrible way they do something because they feel like they're protecting the rest of the organization from its terribleness. And they almost don't want to be saved. What was the conversation like where you walked her through it—you don't have to do this anymore, you could focus on this other thing?
Michael Purifoy
I think in the beginning … she's probably one of about 75% of our workforce I would put them into this bucket, right? They're not just afraid that AI is going to change the way they work. They're afraid that it's going to take their job completely. And so I think it's just sitting down and having that conversation with her and even with others, helping them to understand that you shouldn't look at this as a threat. It's an opportunity, and it's an opportunity for us to put our people in a better position to where they can improve the customer experience and all these repetitive tasks that they honestly hate doing. All they've got to do is shift what they want to do, which is working with the customer. And so there's definitely some resistance there. I think there's a lot in the news right now. And so people are definitely scared of what it's going to do to their job. And if you've been doing the same thing for the last 10 or 15 years, change is hard. And so it's just helping them to see that the change is going to be positive for them, for the bank, and for the customer.
Adam Blue
Yeah. I think it's interesting you bring that up because 75%, that's probably about the same fraction of your employees that in a given day predominantly make granular decisions and execute tasks. And they work for someone and they probably like their manager, but they're not on the handle end of that sword. And so their relationship to the level of change that this technology brings is naturally going to be difficult. And I think the amount of empathy, as leaders and technologists, we can bring to that—there's some humbleness, I think, that's useful. But I don't think you have to worry about that perspective anymore. Because now that they're up for IPO, Dario Amodei and Sam Altman say AI is not going to take away all the white-collar jobs, and we should all just buy shares of their stock and it's going to work out fine. So I think that concern is totally addressed.
So we've talked a little bit about where you guys are in your journey, which I think is fantastic. We talked a little bit about a fantastic real use case for AI. So let's talk a little bit about something that I think is really challenging, which is these notions of governance and compliance and regulatory elements. And I'm going to challenge you here a little bit. Maybe tell me a story in which the introduction of generative AI improves your ability to be compliant and well-regulated and generate the documentation you need for examiners, as opposed to making it harder. Is it possible to tell that story where AI makes it easier and better and more efficient to do those things?
Michael Purifoy
Yeah, I would say so. I think, just take our knowledge base for an example, right? What we were seeing is that we could ask five employees the same question and you might get three different answers to the customer. And so I think, yes, to auditors or different ones too, but just talking about the customer itself, what we realize is that that's not a good thing, right? We have an answer. It's one good answer. And we should be telling that to our customers. We should be consistent there.
And so when we rolled out the knowledge base, there was definitely some resistance there. There was a lack of trust. And so adoption was slow. But once the employees tested it over and over and they could see that the answers were true—we could show them the document it was coming from—they kind of jumped in the deep end. They embraced it. And so I think just being able to be consistent back to what our policies and procedures are, back to the type of answers and what the output back to our customers we want it to be, tells a story that's good. It's good for our customers. And I think regulators see it as a good thing as well.
Adam Blue
Yeah, that makes a lot of sense. And honestly, what the regulators really care about is your answer to the customer, right? It's not the answer to them that matters. They want to know that you behave and you perform consistently, because that's how a bank projects trust into the world, which is, I think, one of your most extraordinary responsibilities. And really the kind of magic thing about community banks in the United States that's so valuable. So that's interesting.
So now we have the shift in culture that's already started. Like you can feel it. And so how do you think at VeraBank about—and I don't know that these are opposing forces, but they're not aligned yet—one of which is we have access to a new technology that is radically different than a lot of technology that we've acquired, because it operates semantically instead of procedurally. It has an element of randomness that is both a source of delight and terror, and it moves faster than we are used to technology moving. And then on the other hand, your customers. They have the same needs that they had yesterday. They need the same things from you. They're counting on you for the same things. And so how do you reconcile the advances in technology and the way you can produce more consistent experiences and better answers and move faster, and then keeping the elements of your culture that I believe and I think you believe really draw your customers to you and are part of your value proposition to them? How can we think about reconciling those two things?
Michael Purifoy
Yeah, I think this is where we're trying to tell our workforce that this is an opportunity for our people to spend more time with our customers and less time doing those repetitive tasks, right? We want the technology to work in the background. I don't know that our customers should even feel it. What they should feel is that we're giving better answers, more consistent answers, that we're moving faster for the customer. And so at the end of the day, it's going to create hopefully an opportunity for us to generate a better customer experience and then more revenue. I mean, we want to use the technology to create capacity within our workforce that they can then turn around and deepen customer relationships and then hopefully produce more growth for the company.
Adam Blue
Interesting. Yeah. I think one of the things I've noticed, because I don't do the kind of agentic coding where you stand like Mickey Mouse in “Fantasia” and wave the wand and then all the buckets and mops do the thing. But I use Claude Code for extending code and writing a lot of unit tests and doing small features. And I like it. I think it's great. What I've noticed is I used to sit down and I would block out an hour where I wasn't talking to an employee. I wasn't talking to a partner, I wasn't talking to a customer. And that was like my hour. And I could just fiddle around in VS Code and it made me happy. I liked that hour. And now, thanks to Claude Code, I can do that work in nine minutes. And that's fantastic, right? And so it means maybe I can do three features in an hour, and then I can use the bathroom, which sometimes I don't get to do during the day if it's real busy and I have back-to-back Zooms.
But the other thing I've noticed—this is just for me personally, but I think it's happening to other people—is that with that increased pace of using the AI tools, the part of the work where you got into the flow of it, and it was like dribbling a basketball or like chopping chives, that almost like physical kinetic part of the work, I'm not getting as much of that. And I'm getting more things done, but man, I get to the end of the day and I might be more tired than I was before. And so I wonder, for employees, how are we going to mediate that?
Because I totally agree with you. I think this notion of let's automate away what's repetitive, let's replace it with agentic technology, let's spend more time with customers. But to be fair, the cognitive and emotional workload of having a conversation with someone is higher than it is sitting down, clicking screens, or writing a document or typing code. And I think we're going to find out. We may find out what the frontier for fully engaged cognitive work is. And I'm not sure that it's eight to nine full hours a day of that level of intensity. And I wonder are you starting to see that with your employee base, or are you still not far enough into the high velocity where that's happening?
Michael Purifoy
I don't think that we're far enough along to really feel that yet. I mean, what our CEO would say is that we don't want to create all this capacity within our production workforce so that they can go play more golf every day. But will they have the opportunity to maybe build relationships playing golf, being able to do that more? Maybe. Yeah. But I think what we're wanting to do is find that happy medium of, again, it not being an excuse to spend more time with customers. Saying, "Well, I can't go do that because I've got all this work that I'm having to do in the background. I'm writing this memo and it's taken me 45 minutes instead of five minutes to really get it where it needs to be." And so, yeah, I think the investment is hopefully turning that workforce into more of those types of activities. But I agree with you. I think there's got to be a happy medium there. It can't be eight hours of just talking all day long.
Adam Blue
No. And yeah, you've got to create some space. Right. And so, as we start to wrap up here, this is something we've talked about on probably two-thirds of our Cut to Context episodes. At some point you've got to start to measure, right? Because we're spending a lot of money. Maybe even more importantly than the hard dollars, we're asking people to change. We're asking them to do their work differently. That's a lot to ask of someone. And so how are you guys approaching measurement, and what have you done around measurement of outcomes that you think is really good, that you could share with us?
Michael Purifoy
Yeah, I think we're in the middle of making sure that we're building out dashboards, KPIs, where we're measuring. Are we, again, starting with the customer experience, improving the customer experience? Are we able to book that loan a little bit faster than we did without the technology? Are we providing consistent answers? Are we seeing that the technology is being adopted appropriately, and at the pace that we expect our workforce to adopt the technology? And so it's not just in one area; it's really across multiple different areas that we're measuring it.
And we're not going to be perfect at it. We're not going to totally get it right. We're definitely going to make some mistakes. There may be an agent that we deploy that we figure out … it's like, we spent time building it and it really didn't move the needle as much as what we thought. And that's OK. We'll back up and go a different route. But what we don't want to do is spend time building an agent and then have it end up on the shelf, or have it really not move the needle, and we just continue on like we have over the last 10 years where we haven't looked at our workflows and we end up in the same spot that we were before.
Adam Blue
Yeah, great. I think that's a fantastic way to think about it. All right. Well, thanks for being on Cut to Context today. Special CONNECT edition, in-person, two-chairs edition of Cut to Context. Fantastic.
So today's takeaway. We like to recommend a piece of art or culture to add. And so there's a fantastic book that I bet a lot of people who are watching have probably never heard of. And it's about women's youth boxing. It's a novel called "Headshot" by a woman named Rita Bullwinkel, and it's a story of 16 amateur female boxers at a single tournament in Reno, Nevada, in the hot, hot summer. And I saw it on a New York Times book list, and the reviews were just glowing. It's not even a very long book—if you're like me and the world has destroyed your attention span—but the author does a fantastic job of individually representing 16 distinct personas and personalities in this book, with almost no dialogue, in the context of these boxing matches between these young women. It's a fantastic piece of literature.
And it just struck me as we were talking that that variance in experience, all those different perspectives people bring—I mean, it's what makes a bank or a software company or any kind of company really fantastic. But if you lose sight of that, the way people respond to things and the fact that everybody gets to be a person and have their own opinion and have their own approach to things, you can really get lost in it.
So thanks again for being on today. Read "Headshot" by Rita Bullwinkel and enjoy the podcast. You can find us wherever your better podcasts are bought and sold. Thanks very much.