AI Basics for Bankers with Ben Udell

Caleb Stevens (00:38.892)
Well, Ben, welcome to the Community Bank podcast. We don’t have many vendors on the show. We have some vendors on the show, but we get pitched a lot by vendors. And it’s not often that we reach out to vendors like yourself personally because of requests that we’ve received and people that have talked about you over the past few months. So welcome to the show. It’s great to have you on.

Ben Udell (01:06.298)
Yeah, thanks, Caleb. I’m excited to be here and I guess honored. I’m a small group of people that get to be on your show on the vendor side.

Caleb Stevens (01:14.062)
We get a lot of requests, a lot of pitches. I probably get two or three PR emails from PR firms emailing us a day or a week or so to have their client on the show. our mutual friend Jack Hubbard, LinkedIn messaged me probably a year ago, maybe two years ago, and he said, you’ve got to have this guy Ben on your show. He is really trailblazing the path for community bankers when it comes to AI.

Ben Udell (01:16.994)
sure.

Ben Udell (01:20.89)
Mm-hmm.

Caleb Stevens (01:42.274)
And you know, even back then, feel like AI was coming on people’s radar, but it didn’t have quite the fervor that it seems to have today. so anyways, as I was thinking about folks we could have on to kind of talk about this topic, you were one that came to mind. And so I guess just to kind of level set things for our listeners, tell us about you yourself, your banking career and what you’re doing now.

Ben Udell (02:03.096)
Yeah, the background is I’m very classically a community banker. So I’ve spent about 30 years in banking and most of the back half of that has been in a community bank. When I first started, were $300 million in three branches. And then when I left, we were around 20 branches in three billion. So very, very classic. I jumped to the vendor side for a little bit. And kind of my origin story is when we were going through a merger with the bank and I am based out of Madison, Wisconsin.

I was leading innovation.

And during a merger and a core conversion, nobody cares. We have fires to put out. And so that’s when ChatGPT came out and I started getting active and engaged and it fit my role and my personality. And all of a sudden I’m traveling around the country talking to other bankers about AI. And one thing led to another. so earlier this year I spun off and I run my own consulting company and I do everything from a lot of enablement and engagement, onsite, private engagements. do a lot of on the speak.

inside the world, teach at Stonier and I teach at lot of the ABA events and I lead AI faculty for GSB Wisconsin. I write for the financial brand on it and kind of get involved in all kinds of areas. So I’m this kind of weird, unique guy who has the banking background and the experience and can actually talk to the bankers and the exec team about activation and enablement and usage. And so it’s been a wild ride and I am all over the country these days.

and it’s been awesome.

Caleb Stevens (03:35.768)
Fantastic. Well, will definitely say anytime a vendor can say they’ve actually worked for a bank, feel like that instantly, you know, automatically raises your street cred in the eyes of the bankers out there. So it’s cool to see that you spent a lot of your career in, you know, normal traditional community banks. And so I guess just to kind of kick off today’s topic, you know, we’ve had some AI discussions for off and on for the past couple of years on this show. But I always like to just

never assume that our audience understands, you know, or has been keeping up with all the things with AI. I’m sure a lot of them have, but, you know, it’s always nice to start with clarity. So I guess for the beginners out there, let’s just start here. You hear a lot of these terms thrown around, generative AI, agentic AI. Talk through, where are we? And I know we’re going to release this episode as quickly as we can after recording, because I know it changes. It changes, feels like day to day, week to week.

Ben Udell (04:31.278)
Ha

Caleb Stevens (04:35.02)
But where we are now, we’re recording this on May 12th for our listeners who are needing some context. Walk us through as we sit now the different types of AI categories out there.

Ben Udell (04:45.57)
Yeah, let me walk through it.

I think very centered around what is applicable for adoption with a community bank right now. You could do way more than what I’m talking about, but I think from what I run into, who I talk to and work with, the mass audience, this is what fits. So from a generative AI standpoint, what we’re really talking about is the Microsoft co-pilots. We’re talking about ChatGPT. Claude has been super popular right now. In fact, if you’re following the news, FIS just entered into some agreements with

Anthropic, which owns Claude to bring a lot of financial enablement to their solution. And then you also have Google Gemini as well. And the whole point is they are very thinking-orientated models. If you want to use the model to write better, to think more clearly, to…

put together content to help be a devil’s advocate for you. are wonderful models for that. And I know we’ll talk more about this, Caleb, but what is I think interesting right now is a lot of executives are not diving into these models and using them to effectively work more productively, to be more creative, to think more critically. And that’s this applicable opportunity that everybody listening has right now.

Now, the other topic that’s been super hot recently is this whole talk about agents. And for anybody that is a kind of an AI or maybe an IT nerd, they’re maybe not gonna like this approach, but this is how I describe it to audiences and who I speak to. From an agent standpoint, if it’s not clear, think of it as your digital employee.

Ben Udell (06:26.946)
What can you have, what would you love to have somebody do for you that you simply don’t have the time, don’t have the energy, don’t have the ability to do it?

take that and with the power of AI, you can create digital employees, you create AI agents. And the easiest one to think through for most of us listening is Microsoft’s Copilot. Microsoft Copilot is this little agent tab and you can go in and imagine for the executives listening, you upload a lot of your writing samples, put it into this little agent so then when you actually write content, your board report, your shareholder report, a letter to the FDIC, something in the community, this

can effectively write for you on your behalf. Now as a human in the loop, you’re not gonna just take that and share it.

But what you are going to do is compress the amount of time down that you’re actually spending typing away on a computer. that’s this like very basic entry point for agents. But if you start thinking about it as a digital employee, all of a sudden people start saying, I would love somebody to do this, this, this, this, and this. And who wouldn’t love extra employees working for them? Now deploy that across the whole mass of your organization. That’s super powerful, especially

especially in the community bank space where we’re all fighting for resources and don’t have enough time and it’s hard to find people. I think those are the biggest things that are opportunities for people right now that if we just get more community bankers engaged in it, they will have the ability to be way more competitive in their marketplace and with larger players and allows them to play up.

Caleb Stevens (08:06.862)
So you’re running workshops all over the country at these banking conferences. I’ve heard people give very high praise reviews of the workshops that you’re running. I’d love to kind of just dive right into it. Talk about like, what are three, four, five like super practical use cases that you’re actively coaching, encouraging, showing community banks right now how AI can really work for their bank? Because I think one of the common themes I hear a lot is,

We think there’s some use cases here, no doubt, but where in the world do we start? Like what’s the crawl, walk, run process we need to take here?

Ben Udell (08:43.8)
Yeah, yeah, it’s a great question. And I think.

The value of a lot of these sessions stems from a couple of things. The first is, I think what we often hear is, you need to use AI with your data. Well, the reality is most people listening are not gonna go out and buy an AI model for their data. They’re not at the point. They don’t understand AI enough. Their organization isn’t using data like that. And so there’s no sense in buying this Ferrari to use for data when you’re not really using a ton of data right now. The other often that comes up is,

risk management and fraud prevention. Well, the reality is in the community bank space, you have your core provider, you have Visa, you have all of these providers that are doing all that for you. You’re not going to spin up a solution that…

will frankly do a better job than what they’re doing. So then I think what happens is all of these bankers are saying, that’s great. What do I do now? How do I use this? And that’s where a lot of my sessions focus on. And for people that are actively using AI, they’re going to kind of have this like no-da moment. But the reality is nobody goes to a concert or a conference and actually shares and provides video examples, which is a big part of what I do. So from the executive side of the world, I’ll give you some really quick hitting ones that are

easy to apply and will save productivity and creativity, et cetera. But one of the things I first talk about with people is a lot of executives, they don’t have time to write. They’re not paid to write. A lot of them are touch typing their way through a computer. And the reality is that the brain power that they need to use to apply to help their bank grow, it’s not a good use of time when you’re typing.

Ben Udell (10:29.634)
And so what I will teach is you can go into Microsoft Word, you can buy an app for this, you can record your sessions and then basically give that transcript to AI and say, write my board report for me, write my shareholder report for me, write my email to the bank, write my script, my presentation, et cetera. We can talk so much faster than we can type and we also don’t have to organize it. And so that’s like automatically this light bulb that starts going off where executives are like,

Huh, if I could save three hours and not type out this board report, that’s an automatic win. Another topic that I work a lot of time on with bankers is just simply communication. and especially in the executive space, they have to work through communication. And I’m gonna call this more specifically like employee topics that just take a lot of time. And a good example I’ll use on stage is I have two executives that do not get along.

One is this personality type, one is this personality type, and they’re opposites. And so the executive can absolutely solve this on his or her own, but the reality is with AI, you can basically describe the situation and get insights on the personality type, reminders of key roadmap issues that could be landmines or opportunities. How am I gonna script and structure this conversation? What are the outcomes I wanna get at? How can I hold people accountable? And it can also challenge your thinking around maybe what you’re gonna say,

is these two executives don’t get along? Who cares? Like in some cases, it just doesn’t matter. Like it’s not affecting the business. They just don’t get along and we want people to get along. And there are other times it’s affecting the business and there needs to be consequences. And so none of the executives listening here are gonna say, know, they can all do that on their own, but they can do that with an AI very, very efficiently. And I’ll give you one other one that I think has a lot of light bulbs that go off for people. And that’s this critical thinking process.

And it’s easy to say that, but what does it mean? And so in the banking side of the world, especially with executives, there’s not often in the community bank space, a lot of executives that have a team of people that can challenge them.

Ben Udell (12:41.208)
that will challenge them. The CFO is a great example of having to deal with the balance sheet and the income statement and the regulatory requirements and completing the call report and thinking critically about Elco. And so you use AI to help them think critically. Give me pros, cons, pluses, minuses. What are challenges? I think this is gonna happen. And so one of the prompts that I like to show with people is I’ll basically take, you can take your bank financials, it’s all public.

You can take the FOMC Fed Minute reports, the Beige Book, Jamie Dimon’s shareholder report, your own analysis for the kind of economic state that we’re in. You can take your strategic report and put all of that into AI and basically start to ask it to be.

a coach, a mentor, what should I do here differently? What could be concerns? Does my strategic plan align with the economic environment and the growth that I’ve seen? And it’s this opportunity to just think critically. And nobody’s saying an executive is gonna take what AI said and you’re gonna apply it blindly.

But if we go back to the concept of there’s not a lot of people in some of these organizations that can really challenge you or are comfortable challenging you, AI is this opportunity to think more critically. And those are like these three like basics from an executive standpoint where automatically they’re saving time and energy, they are being more creative, they’re thinking more critically. And when we get down to spending $30 a month on a license for some of this stuff,

That’s money well spent. And by the way, none of those examples I gave anybody has any PII data or private information in it as well. It doesn’t.

Caleb Stevens (14:20.622)
Yeah. So going back to what you said about writing, one kind of devil’s advocate pushback I’ve heard from folks is to what degree is it a good idea to outsource our thinking? example, for example, writing can be a very clarifying process for your own mind. It helps you understand something complex. If you can’t simplify it, you don’t understand it. Well, if you outsource it to Claude or to chat GPT here, or Copilot or wherever, are you short circuiting

the mental gymnastics that it probably takes to build muscle. On the other side, I’ve also heard the phrase or the saying, well, people thought the calculator would make us dumb too, and yet the calculator actually improved our ability to solve even more complex equations. Where do you sort of see, I guess, the pitfalls there, just from a cognitive capability standpoint, when, you know, it takes a lot to be an executive and there’s a level of thought and intelligence that you got to have and that’s a muscle you develop.

Ben Udell (15:18.008)
Yeah, it’s a great topic. Certainly a very common one to talk through. And let’s touch on one of the very basics, first of all, and then we’ll kind of weave even deeper into that. The first one that comes up and I think it’s…

always interesting when people want to talk about this and that is the concept of work slop effectively they’re taking what AI has and they’re just publishing it and it’s sloppy work. It’s not good work. And I think too when you think through that with AI that’s not an AI issue. That’s a you issue. You didn’t prompt well or dive deeper into AI. You didn’t think as critically and creatively about it as you should which I think Caleb is exactly what you’re talking about right. You’re using AI to have a thoughtful conversation back and forth.

Or you’re just lazy and you don’t fact check, don’t review, don’t review for your brand and you send it out. And whether that’s with AI or you just do sloppy work, either way that’s a you issue. And so I think what’s happening here, when I think through the writing aspect of all of this, yes, and I’m one of these people too, I like to write just because it clarifies my thoughts. But what I also find is,

I don’t want it to write 100 % of what I do and I probably don’t want it to write 80 % of what I want it to do. I want it to write the 50 % of what I want it to do that reserves that other 50 % for my creative thinking, my thoughtfulness, and I think a good use or thought approach to all of this is it’s this kind of dynamic piece of some cases you’re gonna let it think for 90 % and you’re moving on. Other cases.

you’re gonna be at 50 % and that’s okay. And I think that’s where people have to get in and work with these models. And one of the examples I’ll often give is if you treat AI like an intern, like a person, you’re gonna have a dynamic conversation back and forth.

Ben Udell (17:06.656)
Which then starts what people question me on is which model should I use should you Claude or Gemini or Chachi PT? My response is I think it’s like having four interns You get to make a choice and I think back to this writing piece If you can find the model that works well for you or interact with it in a manner that works well you are enhancing the amount of time you’re spending thinking critically rather than Just letting AI work on your bath. I think that’s this kind of big wrap-up key to this is you could absolutely

absolutely dial it in and or mail it in I should say phone it in and just ship it or you can have AI do a Big heavy lift and we could talk a little bit about something like the commercial lending side of that in a second on on how that’s so so valuable like It’s just using it as a thought partner not a site that just does the work for you

Caleb Stevens (17:55.374)
That’s great and I’d love to get back to the commercial lending implications in just a moment. One topic I did want to touch on is, you know, a lot of bankers, one of the things you’ll hear is they’ll say, well, we’re not so sure about the regulatory implications. What are the expectations from our regulators? What does compliance look like? And those are all very valid, very valid questions. At the same time, you know, sometimes it could be sort of a shield for

Ben Udell (18:16.238)
Mm-hmm.

Caleb Stevens (18:22.208)
resistance to change and this is the way we’ve always done it and you know, it’s more of an internal culture issue than it is truly a regulatory hurdle. Where do you sort of see that tension and what are the expectations as at least as they stand now from the regulatory bodies?

Ben Udell (18:37.834)
Yeah, yeah, it’s a great question. Interestingly, I’m a few as maybe being dramatic, but I’m one of the few bankers that’s actually spoken to the FDIC about our AI usage. So before I left the bank, the FDIC, I think it was our safety and soundness exam.

When you’re traveling around the country talking about AI, you can’t escape the conversation. And we had a great discussion. The FTAC was great. And I think what people need to think through with this process is let’s start very foundationally for a second. If I have an intern that’s writing a marketing campaign for me,

I’m gonna take that marketing campaign and I’m gonna run it through my compliance team, my marketing team, my subject matter expert. I’m gonna look for factual information, compliant information.

unbiased information. When we use AI, we have to remember we’re doing exactly the same thing at this point in time. And so I think when people get really worried about that, when we take a step back and say your process is technically still the same, you just have a computer that’s putting together a lot of work for you rather than a human being. And that, think, is a way for us to think more critically about frankly being smart about what’s actually touching the consumer. I think the other thing that

often comes up is they’re worried about some of the chat bots giving bad information. And my response is…

Ben Udell (20:02.904)
Have you met your bankers? You have 24 year old bankers with limited banking experience and they are answering questions based upon a two week training and they’re trying to find responses in SharePoint with five different versions of the same policy. So you can’t tell me that AI is going to do a worse job in answering a lot of these questions. We’re conflating risk here and I’m not dismissing the risk of AI. I just think we’re not thinking

We’re not thinking about those at the same time together. Now, when you actually start getting to some of this data work and some of this other stuff, I 100 % agree. I think that’s more challenging, but for right now, where I see most community bankers, they are not fully deploying generative AI.

They’re not actively deploying PII and they could and should be. So let’s get through those gates to understand before we actually get into some of these models with some of the actual vendors. The vendors are doing great work. But I will tell you, think the funny thing to me is people are uncomfortable with adding PII through Microsoft Copilot right now. But then I’ll hear them turn around and they’ll talk about, I’m talking to this vendor about having this on-prem SLM, small language model.

And I think I might want to go with them like wait a minute you you just you even did not hear about this vendor until a week ago at a conference you’ve been working with Microsoft for 40 years like let’s let’s be real smart about our risk management and vendor management processes for a lot of these things so It’s it’s it’s we’re not understanding the risks super well at times and I would wrap that up by saying I think in a lot of cases if people are not deploying out well they generally have actually not actually done the vendor management for the risk management

They just see the headlines and say, we’re not there yet. We want to be a follower rather than I think the executives should be saying whether we’re going to apply AI or not or PI right now or not. I need these different groups to be giving me a quarterly review of the models, the risks, what’s the trust policy look like? What are issues they’ve seen? How could we deploy it? Because if you don’t do that, how do you actually know how to move forward? And by the way,

Ben Udell (22:11.458)
you’re now behind because people aren’t actually doing the research and understanding a lot of it. That’s, that’s, that if I had a pet peeve and a soapbox I would get on is do the vendor management, do the risk analysis, do the education, even if you’re not gonna turn AI on. But we’re at this tipping point now where people are gonna start losing talent if they are not doing more with AI. So that’s a problem and a risk that people are not actively talking about.

Caleb Stevens (22:30.68)
Yeah.

Caleb Stevens (22:34.382)
Well, you’re example of having a hunt through SharePoint to find a policy or a document. I think we’ve all been there in the past, so that’s a great example of an easy way that AI can save us all time and frustration as well. Let’s go back to the commercial lending piece. So we’ve talked to a lot of commercial lenders out there and you know one of the key themes you hear and rightly so is we’re all about relationships. Relationships. We go take our clients to play golf. We get to know them. We go to church with them. We’re in their offices. We’ve

Ben Udell (22:44.58)
Mm-hmm. Mm-hmm.

Caleb Stevens (23:03.726)
Our kids play little league sports together. We are a relationship. So, you know, why do we need a CRM? I hear that a lot. You know, I don’t know. I don’t need to be spending my time in a CRM. I need to be out in the field calling on on prospects. And there’s there’s some merit to that. But I often think that maybe sometimes there’s just maybe not quite a full understanding of all the ways that AI and the emerging technologies out there could help.

Ben Udell (23:11.822)
Sure.

Caleb Stevens (23:30.752)
even a commercial under to be more effective. what’s your perspective on that?

Ben Udell (23:35.406)
Yeah, first thing I would say is the relationship aspect absolutely doesn’t change with AI. And I think we need to appreciate the fact that it’s not going to. And I would argue through some examples, it’s going to improve.

I think what often happens though is that the larger players are using AI not to necessarily replace the relationship. I mean, yes, but then I would argue the customer probably doesn’t care as much. But what I would argue is they are compressing the amount of time to make a decision, to do the analysis, to do the research. And so if the average commercial loan from start to finish is maybe gonna take three to four weeks right now,

What I would argue, how much of that three to four weeks is literally somebody sitting down at a computer and reviewing the loan, doing the critical thinking? I would argue that’s hours worth of work, if not less. But we just said it just took three to four weeks.

what gives here. And so when we think through the commercial lending, what I really want people to kind of think through is how does AI support compressing the amount of time without losing the relationship? Well, one good example is just simply doing research.

Tell me about this provider, this vendor, this commercial real estate company, this construction company, this architect, doing all of the resource, citing the sources, packaging together for me, really like a white paper about who they are. So now I have some credibility to have a discussion. And then I can say, put together a discovery call agenda that I wanna ask a lot of questions about. And so it starts to formulate that.

Ben Udell (25:16.184)
Then you might say, I can record that conversation, now I have that recorded conversation to be able to take it and say, help me understand what are the hot points for this individual, put together a next step. I also have their financial statements. Let’s do the analysis of the financial statements and let’s compare the financial statements versus my loan policy versus standard ratios and expectations for the industry that this person is. Graph it, chart it, highlight it.

questions.

What are the questions I need to ask for follow up now because it might be out of loan policy or it might be in some murky gray area. And the point is you start building all of this together and we’re not removing the human from any of this. What we’re doing is we’re accelerating the busy work so we can actually think more creatively, more critically about what they’re doing. And the wrap up to this is everybody ends up putting together, you know, 10, 15, 20 some page loan packet that they’re gonna take to loan officer committee, right?

Great. Have AI put all that together. Here’s all of the documentation that I have on this customer. Here’s an example of what I want to build. Let’s build all of this together. And all of a sudden, if you have all that documentation with an example, you can build your first draft in 10 minutes. Or.

You could do it old school and you could wait for somebody to type all that together and it’s going to take three days because it’s basically a term paper. And the fun one I would add to that is I’m going to go to officer loan committee. have a curmudgeon banker who’s going to ask me a lot of questions when they seem out of policy from my loan policy. What are the three topics that are out of loan policy or borderline? How do I address them? How are we mitigating that risk? How should I present this in the meeting? How do I want to overcome those concerns? What concessions are we

Ben Udell (27:04.143)
making with the loan to account for that added risk. And now all of a sudden this banker goes into loan committee well prepared for those discussions, challenges, et cetera. And one thing you could easily do is you could take all of the questions that loan committee generates and basically upload those and say, are the kind of questions that I get on my loans. Help me.

prepare and prep. And so now, not only have we compressed this amount of time, we’ve thought through it more critically and we’re more prepared to have, frankly, to be challenged on And we should be challenged on the loan, right? These are important opportunities to think critically about. But I’ve thought better about it ahead of time and I’m prepared for that actual meeting. You can do that really, really fast. Really, really fast.

Caleb Stevens (27:48.463)
That’s great. one of the common things you hear in the commercial banking world is the importance of responsiveness and speed when you have a client in front of your request from a prospect for a loan or a new transaction account. And the faster you can be getting papers turned around, credit spreads turned around, all the things that are needed to make a decision, the Prep for Loan Committee, all that stuff’s great.

Ben Udell (27:58.318)
Yeah.

Ben Udell (28:07.608)
Yeah.

Ben Udell (28:10.906)
Yeah, I think.

I think to that point, like it’s easy to basically drop the financials in and your loan policy in and some basic ratios and say, give me 10 areas of success you want to really reinforce what they’re doing and pat them on the back, right? Good relationship building. What are the 10 things that are concerning? And you’re not gonna use all 10, but you can then use your critical thinking to say, these three are really powerful and I’m gonna have to know what they are. And I think the other thing we lose sight of on this relationship piece is a lot of these customers, they’re using AI themselves. So they’re very likely, some of

are like, why does it take three to four weeks? You have my financials. Make the decision. Or how many of them are rushing for the right reasons or they want to get it done and how much extra time

Are they going, how much extra time can you take before they say, I’m gonna go with a different player that’s using AI? Or, I am doing the bottom line here and for an extra quarter point or five basis points, the relationship doesn’t matter to me. That’s a challenge. We think relationships are really important and they absolutely are, but for some people that pressure gauge of.

Caleb Stevens (29:25.614)
Well, when your competitors getting stuff turned around a week before you can, I mean, that’s, I would say that’s an improvement upon the relationship. You’re making an investment in improving the relationship with your prospect or client because you’re more responsive and you’re faster.

Ben Udell (29:28.602)
Yeah.

Ben Udell (29:35.769)
Yeah.

I would honestly wonder if you could turn it around in two weeks rather than four weeks or a week instead of four weeks and the competitor is basically over pricing by a quarter point, like are they just gonna sign it because they’re ready to move on and it seems fair or are they gonna wait two to three weeks to save a quarter point that they, I mean, they don’t know if they will or won’t but I think the reality is there’s some pricing power to move swiftly yet responsibly.

Caleb Stevens (29:52.29)
Right. Right.

Caleb Stevens (30:03.79)
Yeah, that’s good. Well, Ben, as we land the plane here, I guess one maybe last topic I’d love to get your thoughts on is in the background of this conversation. I’m sure a lot of folks are listening, worried about the implications of how AI might replace their jobs over time. You hear a lot of doom and gloom, you read a lot of doom and gloom articles, and it doesn’t take very long to scroll on X to have a lot of fear flood your mind when your algorithm gets

Ben Udell (30:30.755)
Mm-hmm.

Caleb Stevens (30:33.58)
you know, to get kind of apocalyptic with AI. So I guess any advice for, you know, the mid-level folks at a bank, maybe they’re in deposit operations, loan operations, credit, and they’re trying to find ways to sort of AI-proof their career. Where would you counsel them in that area?

Ben Udell (30:49.402)
Yeah, I would counsel them with two things. And the first is, let’s think a little higher, bigger picture for a second here. One of the things when I do speaking engagements is I’ll ask how many of you are actively using AI at your bank right now? Let’s say half the bank raises their hand. Half the room raises their hand. Then I ask how many of you would be willing to move to another bank if they don’t offer the ability to use AI and everybody puts their hand down? And so I think one of the un…

kind of spoken risks right now is not the PII and the data, it’s the talent management. If you have high performers and you’re saying you need to do eight hours of work in eight hours or the bank across the street is gonna let you do eight hours worth of work in four hours, A, you’re gonna have people that start to unplug because there’s a real value to work-life balance being transformational. If you wanna be a community banker five years from now, you probably have to do AI, right? Like at the end of the day, do you wanna be the last person to turn the lights off? Probably not.

think that’s a big part of it. I think with that in mind then, with the downstream piece for that is if you haven’t put somebody in charge of AI, you haven’t rewritten job descriptions, you have no goals. What that also says to me is on January 1st, 2027, if you’ve made zero progress on AI, mission accomplished. No goals, no job description.

no opportunities, you would expect no progress. And that’s a statement that I think a lot of people are seeing right now where they’re like AI is a part of that back to this job conversation. So then when I just kind of swoop in and think through your commentary on the job side, I think it’s incumbent for everybody right now to figure out how they’re going to use AI to work more productively, smarter, more creatively. The group listening and their people have really great customer relationship skills. If we believe a community bank builds relationships,

then let’s use AI to give them more time to build relationships. If somebody is doing work that a computer can do at some point in time, we have to say, a computer can do that. How do I redefine that role? And I think personally in the community bank space right now, community bankers are struggling to find people and they’re also resistant often to bring people in from non-banking roles.

Ben Udell (33:02.742)
AI as a solution for all of that at this level and at the FTE counts that we actually have. So I wouldn’t dismiss that anxiety, but I do think from a leadership standpoint, they are signaling what is and is not acceptable. They’re encouraging responsible usage. They don’t have to lead by necessarily doing AI, but they do have to lead by being very proactive and engaging and supporting of their people as well.

And I’ll wrap up with this, Caleb. The other question I ask is how many of you have received actual AI training? Less than 5 % of people raised their hand, which is great business for me, let me tell ya. But how do you make people feel comfortable? You provide them training on how they can use it effectively to be better employees, not by leaving it shrouded in mystery. And frankly, many of them are uncomfortable to use AI because they don’t appreciate how hard they can step on the gas.

Caleb Stevens (33:38.19)
Yeah. Yeah. Yeah. Right.

Caleb Stevens (33:56.771)
Yeah, that makes a lot of sense. And that’s a great call out for the listeners is, one, do you have a roadmap and a plan and a strategy? And two, are you implementing that in a way that your team members can actually grab onto it and learn and develop their careers through that? Ben, if folks want to reach you, if they want to find out all the resources that you offer for the banks out there, how can they find you?

Ben Udell (34:06.873)
Yeah.

Ben Udell (34:15.874)
Yeah, easiest way is certainly on LinkedIn. I’m easy to find under obviously Ben Udell. My website benudell.ai and certainly they can find me there but also email me at ben at udellconsulting.com. I do a lot of kind of public, private, virtual, on-site engagements to help not only the entire organization but executives.

Move on AI. I’m at the Oklahoma users group right now for community bankers. I’m like, I bet everybody in this room wants to be a community banker five years from now. And we know AI is coming for us. So let’s help our teams grow and spend more time in the community so we can be community bankers five years from now. That’s an important message.

Caleb Stevens (34:58.254)
Absolutely. Well, Ben, thanks for the time. Thanks for sharing your thoughts with our listeners. And we look forward to keeping up with you on LinkedIn and following all the resources that you put out there for bankers. So thanks for coming on.

Ben Udell (35:00.346)
Yeah, thanks Caleb.

Ben Udell (35:08.57)
Sounds great. Yeah, thanks for having me, Caleb.