AI for Lenders: Smarter Tools for Sales Success

Nancy Ozawa (00:06): Welcome to Banking Out Loud. I'm Nancy Ozawa. I'm chief marketing officer here at PCBB, and I'm also one of the hosts of this podcast. Today, we're gonna be talking about something that's getting a lot of attention, AI. But we really wanna focus on the practical pieces of AI. We wanna focus on how AI is actually changing sales in banking. Now, you think about everything that goes into a sales role. That could be prospecting. It could be preparing for meetings. It can be follow-up, staying organized, all of those tasks that a salesperson needs to do. AI is starting to act like a sales assistant, sitting alongside of you, helping you move faster and showing up better prepared. We're gonna look at what's possible today, and we wanna focus on the sales needs of lenders and other customer-facing teams in your institution. Where are those tools really making a difference? (00:57): Were they not quite there yet? We also wanna talk about the human side. How do you use AI to improve how you show up for customers? Whether that's practicing for conversations or spotting new opportunities. We also wanna touch on some of the risks. We wanna be honest about that, where AI could go wrong and what mistakes you might want to avoid. So if you've been wondering how to use this shiny new tool, but wanna be more effective in your sales role than this conversation's gonna be for you. So to help me unpack this episode, I'm joined by my co-host, Joellen McKinley, PCBB's SVP and National Business Development Manager. She brings over 35 years of banking and capital market experience working with community financial institutions. Joellen, it's great to have you on this conversation with me. Jo Ellen McKinley (01:46): Thank you, Nancy. It's great to be here. This topic is one I'm genuinely passionate about, so I'm really looking forward to a practical conversation today. Nancy Ozawa (01:54): Great. And we're also joined by our guest, Craig Kirkpatrick, co-founder of Optimal Advisor AI, the AI learning partner for financial services firms. Craig has over 25 years of experience in financial services, and lately has been training financial professionals on how to use AI. And one interesting fact, Craig, that I was looking up on you is that you have over 5,000 hours and over 15,000 prompts on AI. So you've really tested out these financial use cases. So, Craig, welcome. Craig Kirkpatrick (02:26): Welcome. Thank you, Nancy, and thank you, Joelle. And excited to talk about this, uh, shiny AI toy that everybody's talking about, as you mentioned. Nancy Ozawa (02:35): Agreed. So can you start with just sharing a little bit more about your background with our listeners, especially your journey into AI and especially working with financial professionals in that more sales capacity? Craig Kirkpatrick (02:48): Yeah. So I've been in financial services for the last 25 plus years, uh, working directly with financial professionals. Four years ago, after selling my company, actually, it was the week that ChatGPT came out, uh, and began, as a consultant, just began playing around with AI, um, and seeing its use cases and thought, boy, there's an opportunity to apply it to the industry that I've known so much about. Um, so myself and a co-founder, we started Optimal Advisor AI. We're, we're sitting here today working with Fortune 500 companies, banks, custodians in the US and around the world. And it's really been one where we're the AI learning partner for these companies. So we help firms, uh, accelerate effective AI adoption. So we're simply trainers. We're helping folks use the tools they already have, like a Copilot or a ChatGPT. Nancy Ozawa (03:42): And our paths cross, because earlier this year, Joellen, you brought Craig into PCBB and helped us conduct, uh, a really interesting, comprehensive training for our own PCBB field reps. Um, and Joellen, I don't know about your 5,000 hours of, uh, AI, but I know you are also an avid user of AI in business development. So do you wanna touch on maybe some of the surprises you learned when you started this training with Craig? Jo Ellen McKinley (04:07): Well, one of the reasons we brought Craig on board was because as community bankers, uh, we hear every day that our bankers are under pressure to do more with less, and AI is really becoming an answer to that particular task. So PCB's goal was to educate our team to leverage AI more efficiently when we're targeting new prospects or trying to deepen our current relationships. So I'll tell you, the training really did surface some genuine surprises. Um, one of the biggest ones for me was discovering that we had an AI super user on our team. This person was already leveraging AI to prospect for new clients and was finding creative ways to deepen our strategic contributions to existing clients. So it was both impressive and eye-opening at the same time. It told us that the appetite for AI was already there on our team. We just needed to give the entire team the framework and the confidence to sort of build it out and follow suit. (05:07): Um, on a personal level, one of the most meaningful things that I took away from Craig's training was the importance of customizing the settings within whatever AI model you choose to use. When you take the time to define your role and your overall objectives within those settings, the responses you get back are really filtered through the lens of who you are within your organization and what you're trying to accomplish, and that kind of changes everything. It's the difference between a generic answer and one that actually speaks to your world as a banker. So Craig, let's ground this in some reality. Our recent ABA-sponsored innovation survey showed the top use cases right now for community financial institutions when it comes to AI are compliance monitoring, process automation, and fraud detection. But what's actually possible with AI today in a sales context for bankers, whether they be lenders or other customer-facing team members? (06:05): Can you walk us through a few examples across prospecting, business development, communication, and the like? Craig Kirkpatrick (06:11): Yeah, sure. In fact, I'll share a couple examples of that, but I wanna take one step back on just what's possible for your community financial institutions. And as, as you're listening to this, and as you're thinking about your job and what you do, there's three areas where AI can help a lot, and I'll show you some use cases. First one is just around rapidly analyzing information. So if you just think about all of the information you consume on a daily, weekly, monthly basis, you know, uh, reports and emails and proposals, AI can help you do that faster to make better informed decisions. So it can consume information. The next one is just improved communication. This is probably the highest use case within a community financial institution's work here is how can I write better emails? How can I communicate better? So there's a marketing biz develop. (07:02): We'll talk about that. Finally, and the last use case, again, as you're thinking about your own role, I don't see this as much, but using it as a strategic partner, using it as a brainstorming partner. I'm trying to solve a problem. I'm trying to figure out in my marketplace where the high growth areas, for example. So those are the three areas. So let, let me give you a couple specific use cases. Let's just say a new regulation comes out. Normally, I'm sure you have a process to consume that information, but using AI as a strategic partner to consume that information provide you with some key takeaways. And even providing a memo to your management team, for example, on the possible implications can really accelerate the speed with which you can, uh, consume that information. So just digesting complicated regulations or digesting complicated information can be helpful. (07:52): On the business development front, again, this is where we're seeing a real impact. Let's say, for example, you have a new client that's gonna come into you. Maybe they're a CFO of, I don't know, a home builder in your local market. How can you use AI to be better prepared for that meeting? Mm-hmm. And whether you're using ChatGPT or a Copilot or Perplexity, so forth. By the way, I'll just share, and, uh, many of the financial organizations seem to fall into the Copilot area because of some of the compliance issues. But the first thing you can do to prepare for that meeting is have AI do a profile on that executive. "Hey, go ahead and read everything you know about this individual here and provide me a summary of that individual. "Right after that, you can then tell it to go to the website of, of that CFO's company and do a data dump on everything you know about that particular company. (08:43): You might not know about the home building business, and you can ask AI to support you in getting that knowledge. So now you have the profile on the individual, you got the profile on the company. Um, then you can ask the question, um, they're coming into my office to talk about different products. From their perspective, what kind of questions might they have of me? Or I could even insert my agenda that I'm planning to use in that meeting. From their perspective, how might they view this agenda? Or what questions might they have for me? So this is a simple use case of getting intelligence and having AI act like the CFO and providing you information so you can be better prepared. So, Joanne, this is, so one use case on the business developer. Pretty simple to do, but it can be pretty powerful.... Jo Ellen McKinley (09:32): And pretty time-saving too, because if you tried to do all that by hand, um, speaking as a former person who did that, it's hours that you're saving. So - Craig Kirkpatrick (09:40): Yeah. Jo Ellen McKinley (09:40): Bravo. And let's look at the crystal ball for just a minute. If you ha - if you looked into your crystal ball and had to imagine five years from now how AI is gonna change the role of business development, what do you foresee? Craig Kirkpatrick (09:54): Yeah. This is a tricky one, of course, to, to, to answer. But one thing is clear, that the information gap between you and your client is closing. Mm-hmm. Right? It's gonna be no longer who has the most information. It's really gonna be the judgment on that information that happens. Because see, in five years with generative AI, that CFO from that company is gonna have a whole bunch of information already about your products, about your pricing, about your competitor's pricing. And so the information gap closes. So if you're just someone giving terms and information, that's gonna be problematic. Where the human comes in, and where I think this is a really, uh, place to excel, is what is the judgment on that information? And that's where years of experience matters. When I think about your community, financial institutions, the level of knowledge that that individual will have on the client's business will be accelerate. (10:52): Your client will know more about you, you will know more about your client. And so the kinds of questions that will come up, I think, in those meetings will, will be more judgment oriented, will be less about the product and the information. And then I still think human connection is gonna be one of the most important skills here, uh, in a world of AI answers. So, you know, again, it's hard to know what will happen, uh, you know, there, but we do know this information gap will close. And then what do you do? Jo Ellen McKinley (11:20): You're absolutely right. I, I agree with you that we're going from a period of where information is nice to have to where it's table stakes. It, it's - Craig Kirkpatrick (11:29): Yes. Jo Ellen McKinley (11:29): Just prerequisite when you walk in the door. Craig Kirkpatrick (11:32): Yep. Jo Ellen McKinley (11:32): So you mentioned Copilot. There are a lot of AI tools out there, a lot of shiny objects to choose from. How should bankers think about which platforms to use for which of their sales activities? Um, you mentioned Copilot. Are there others out there that kinda really shine as well? Craig Kirkpatrick (11:47): There, there are, but it is interesting. I'm finding with institutions, there's a compliance decision made first. Which model are we most comfortable from a compliance standpoint? And by the way, I found whether it's Copilot, whether it's ChatGPT or whether it's Perplexity, all of them are gonna provide a, you know, a fairly similar, you know, output in many cases. What. John, you and I have talked a lot about this, that the new skill that we will need to learn is the skill called prompting or giving instructions. If you're able to instruct that model to deliver the output you want, that will be the key skill. We've given instructions our whole lives, for sure. We're just, are not used to giving instructions to a computer. So I think I see compliance l- leading the charge there. I do find that many of them are on the copilot because you're in the M365 compliance area and they feel safe with client data. (12:42): But again, it's more about your ability to get the answer out of the model almost rather than the model itself. Nancy Ozawa (12:49): Craig, let me just, uh, ask you a question about that promptings. There was some tricks, there's techniques. Why don't we delve a little bit deeper into that particular piece? Do you wanna share some lessons? Craig Kirkpatrick (12:59): Yeah. The number one mistake that I see people make using, um, the AI models is treating it like Google. See, with Google, we ask questions. And so with AI, we need to give instructions. And so you have this powerful assistant. In fact, one of the messages, everyone on this call, everyone listening, you have access to every expert in the world to draw from. You just need to know how to engage that individual. So, Nancy, the more specific you can be with instructions, we talked about that in the training. For example, AI doesn't know if you're a gardener or a chemist. So you need to provide context. That's a big one so you can give more information. Um, I've seen people begin to use AI and they get frustrated after a couple weeks because the outputs aren't great. You know, it doesn't sound like me. And one piece of advice is if the output is not coming back correctly, you have to be relentless with providing instructions around.... (13:58): I want it to be professional and friendly and no jargon. Uh, so you have to be, you have to instruct it. Joellen said at the outset, everybody has a custom instructions in their AI model. If you go in there and change that, you can improve the outputs. Let me mention one other thing real quick about AI models. I don't know if everyone's noticed this, that's using AI, but they tend to happy talk you. Oh my gosh, every question you ask, amazing. Yeah. That was amazing. You're a wonderful person. A great idea. All of the models do this because they want you to stay engaged. Yeah. And I can just say that you just need to be relentless, uh, about instructing your AI model to not do that. Be candid. Don't happy talk me. Uh, I, I shared with you - Well, I like the Nancy Ozawa (14:43): Happy talk, but yeah, I love changing. Craig Kirkpatrick (14:45): You're great. But listen, when you're doing critical thinking skills, when you're doing something that's really important, when you're researching an industry that you wanna make a decision on, there's. I shared this with you guys before, that there's a Stanford professor when he's doing critical thinking with AI prompting. He tells it to act like a Cold War Russian gymnast judge to, to just critique his answer. So listen, the prompting side of this is a big deal, getting improved answers, and I just give specific instructions, give context. Um, and if you don't like the output, you just need to push back. It's literally three or four iterations down where you can begin to, you know, improve. Nancy Ozawa (15:21): I also learned that not only is it the context, but you have to tell it what the role is, who the audience is. And then the other one I've been doing is ask it to interview me. I'm gonna let you ask up to three qualifying questions one at a time before you answer. And then it asks me questions and I'll realize I didn't give it enough context. But that interview task allows it to get more context out of my head. And I find that tailors the prompts. Craig Kirkpatrick (15:46): You know what, Nancy, that's such a great point. You give a prompt, and unbeknownst to you're leaving out a bunch of stuff. So if you simply added, "Before you answer, do you have any questions for me?" And then AI will come back and so forth. So it, listen, AI is saving you time, but you wanna save time using AI. We just helped a financial advisor analyze a 1100-page tax code, uh, because they had a question from a client and we instructed it to act like a US tax attorney. Uh, act like a CPA who works with small businesses. And it turns out when you assign a role to AI, it focuses the answer and it can improve. Jo Ellen McKinley (16:29): I've also started asking it, what did I forget to ask you? What am I, what am I omitting in my question? And sometimes it'll come back with a whole list of things that, uh, never even occurred to me. Nancy Ozawa (16:40): Yeah. Sometimes it'll come back and say, "You gave me two messages, but which one's more important?" Jo Ellen McKinley (16:45): Right. Nancy Ozawa (16:45): That really helps too. So it's a conversation that you're really having with the system, not just tell it what to do. Jo Ellen McKinley (16:53): So, Craig, one of the things we discovered, as I mentioned, is that we had a super user on our team that we weren't aware of. And I suspect that's true at a lot of community financial institutions. There are people there behind the scenes that may not necessarily be making themselves known. How should financial institutions identify and leverage into these internal champions to help their whole team - Craig Kirkpatrick (17:17): Yeah. Jo Ellen McKinley (17:17): Benefit? Craig Kirkpatrick (17:18): Yeah. So it's obviously about making AI training, uh, a commitment within the organization to unpack this. But here's what we noticed within financial institutions and advisory firms. Or let's say maybe your company has 300 people that work there. Probably 25% of those folks really haven't touched AI. They know they should, but they just haven't, you know, they don't even know where the Copilot button is. There's another 60% or so that are using it to write emails and do some items. And then there is that five or 10% ninja group who have leaned in and, you know, are really able to use it. To identify those people and have them help the other group can be a really important time saver. There was just a study that came out, uh, 5,000 individuals last year, and they looked at all their responses. And they found that about only 10% were really proficient. (18:13): In other words, you got the answer that you should. They just u- updated it two months ago. So same 5,000. And it, and proficiency last year was 10%. It turned out proficiency dropped down, dropped to 3%. So usage went up, but proficiency went down, which doesn't make sense. And the reason is what's happened in the last 90 days is the models have gotten so good. And people are still using it the way they've been using it. So it's great people are adopting it. It's great that they're using it. There's this interesting opportunity to, to leverage the new information from the models and, and increase in the proficiency. Jo Ellen McKinley (18:53): Yeah. It seems like they're learning something new every day. So you're right. Yep. Craig Kirkpatrick (18:57): Let me just say, there is no AI or technology experience required at all to learn this. And you said earlier, just having a conversation. So demystifying it to those people who haven't touched it before, just to show them, just go in and have a conversation, uh, is really the first step for adoption. The other thing we've been telling everybody to take a yellow stickum and put it in the bottom of their computer and just say, "Ask AI." It's just a daily reminder of like, "Oh, wait, I have access to every expert in the world." And they, they're not remembering that they can leverage and lean in. So it's just learning a new habit. It's learning a new technique and it's learning a new way, uh, to engage. But many people on this call were around before computers showed up and they figured that out. So they're gonna figure this out too. Jo Ellen McKinley (19:44): Well, speaking of new ways to engage with AI, um, one use case that kinda stands out to us is a coaching and role-play tool. How can a commercial lender, for example, use AI to practice conversations, get feedback on how they come across, or prepare for specific client scenarios? Craig Kirkpatrick (20:03): At the start, I talked about using AI as a thought partner or a brainstorming partner. And then, Nancy, you talked about the importance of assigning a role to AI. So, um, act like a sales business coach. And then you just provide context. I'm in sales development and I want you to act like a sales coach here. And, um, go ahead and ask me some questions and then I'll respond and you give your feedback. So again, you can tell it to take on a particular persona and then begin to go back and forth with AI. We do this with a number of sales groups where we take their product and we tell AI to act like a grumpy customer. And, and so again, everybody on the call, think about, you know, your clients, who they are. And you have a product. And so you put that into AI and he's like, "Here's my customer. (20:55): I want you to act like a grumpy customer." And they'll come back with questions and objections. And so, Jolan, to, to use it as a sales coach. Now, a lot of times, the first round is not gonna be good. The sales coach isn't too good. The sales coach is too happy. The sales coach is whatever. But this is where you go in and say, "No, no, no. I want you to be, you know, the Tony Robbins of sales or, or whatever you assign it. You know, be, be great. No happy talk. I want this to be realistic." So it takes a bit to be able to talk and instruct AI that way. We also talk about this. It turns out that the older you are and the more life and business experience you have, you have a huge advantage using AI because you have life experience. (21:42): So you can see an output and you'll know if that's right or not. You'll know whether to push back. I'm training some recent college grads on, on AI, and it's, it's tricky because they, they have no perspective of what the output is. It's the first time in human history where technology's come out and it serves the older folks with more business experience. And so to your answer about the role play in the coaching, you'll know what a good answer looks like or not, and you'll just continue to critique till you can get the frame and your AI assistant coach to act the way you want.... Jo Ellen McKinley (22:14): Yeah, it's a great way to anticipate potential hurdles before you walk into a meeting. And like you said, the grumpier, the better because then you're preparing for worst case scenario. Craig Kirkpatrick (22:24): Some teams are sharing prompts within the team itself for the training part of it. Nancy Ozawa (22:30): That's a really good point, is when you do get a good prompt is to share it with some others that are like you so that each one of you is helping each other. Jo Ellen McKinley (22:38): So Craig, one of the biggest concerns that we're hearing from community financial institutions is that AI-generated content can sometimes feel very generic or, or impersonal. Um, in a relationship-driven business like banking, um, that can be an issue. What are your best strategies for taking that AI-generated sales suggestion or communication draft and making it sound less like a machine and more like authentic - Yeah. Personalized voice? Craig Kirkpatrick (23:05): This is a big deal. How can you incorporate AI to accelerate and grow, but don't lose the personal touch? Right? Which is what we want. We want the human side of this. So these items are, are manageable. It gets back to instructing the AI assistant to talk like a human, to be empathetic. Uh, one of the strategies we'll teach our financial clients is how to get AI to talk more like me. So we'll have them insert something they've written and then we'll have AI try to mimic that. Um, the other big compliance concern that we've got with AI, and this is the big deal, is AI will provide an answer that it's thinks is right and sounds right, but it's not. So the folks listening to this call, one of your biggest risks is relying on information that is incorrect. So how can you verify AI outputs? (23:59): You can do things like defend your answer or cite your sources. Uh, you don't believe everything you read on Google and you can't do it here. Uh, we instruct our clients to verify AI outputs. Uh, so as you're using AI to communicate locally to clients, you have to be relentless with AI. No jargon. Be professional. Be friendly. Talk like a human. Joellen, you talked about your custom instructions. You can go in at the outset and set up AI to talk a little bit more human. And I, I think this is a really important point. And by the way, even after you do this, 30% of the time it'll still give you jargon and, and so forth. So you own, you own what you communicate. You have to trust but verify. I see LinkedIn posts every day that are just, you can tell AI wrote. (24:51): And it's, and it's kind of lazy and you run the risk of, uh, just being negative with your clients. So it's being aware of it and being conscious and know that you are not gonna hit send on something till you know that it sounds like something you wrote. Jo Ellen McKinley (25:06): And sometimes AI will give me like three different choices to, to sound how I want to sound. Professional, conversational, concise, et cetera. And you just kind of pick the one that fits where you're going. Craig Kirkpatrick (25:19): Yeah. Nancy Ozawa (25:19): And then you can tell it, "Hey, I always wanna be like option B." And so it puts it into its memory bank to be more options like B. Craig Kirkpatrick (25:27): Right. Nancy Ozawa (25:28): But Craig, let me stay in that risk piece because, you know, there is risks of AI and you just mentioned them. Uh, kind of some of the, the information may be too high level. It may not be as personalized. What are other risks that lenders or other sales-focused people should be aware of when they're using AI? Craig Kirkpatrick (25:46): Depending on the kind of model that they have, client confidential information you just can't put into an AI model. And so you wouldn't put your client financial statement in Google, right? So don't do it here. You, your view should be these are public websites. So Copilot and some of the enterprise solutions can protect your data. But the risk is that this information you're getting or taking just isn't correct. So that, that's where we spend a lot of time just trying to help people make sure that they verify the information. It is interesting when you attach a document and say, "Please summarize." It turns out the hallucination factor is much less because it has the document, it reads it, and it can, can provide output. Imagine you have an assistant that's just handed you this thing and you would say, "Listen, are you, are you sure that these, the, the formulas are right? (26:34): Have you double checked those? Can you verify that this is right?" So it's, it's getting the output and then it's, it's stress testing and making sure that you've asked it a number of different ways to verify that output. Nancy Ozawa (26:47): That's a good point because if it's a person, I would always say, "Did you check your work?" But there's a sense if it's a computer or technology, you think it's gonna be smarter than a human. You may not ask it, but you have to treat it just like a human on your team and ask the same kind of questions. Yeah. By Craig Kirkpatrick (27:01): The way, I'll give you a quick, uh, another prompting technique. Yeah. So imagine that assistant hands you this document and you said, "Hey, can you reflect on your output and provide me any improvements?" And then they go back. If you do that with an AI model, almost 100% of the time, you'll get a better answer. Right. So double check, reflect, tell me where you might be wrong. You asked about what's gonna be in five years - uh-huh. And what we need to do and the critical thinking skills we need to have. Your ability to stress test every document you see, where did you get it? What are the biases? Is it true? That, that effort, I think, will be a critical part of people's consumption of data, uh, in a world of AI-generated content. So stress testing the content. Is it right? What's the opposite? Will be a really important skill. Nancy Ozawa (27:52): I've also found, and maybe it's just the AI that I use, it prefers PDFs over Word docs. Craig Kirkpatrick (27:58): Completely. Yes, you're absolutely right. Uh, or if you. Let's just say you have a 300-word article you wanted to read. You could copy and put it into the text, but you wanna put it in a PDF, so it's concentrated its focus. Yeah. So, um, yeah, PDFs are better than, better than Word documents. Nancy Ozawa (28:17): Right. We talked about risks, and I think we have to touch on people should be aware of what their AI policy is as well. I, I know we had a recent AI episode, and that person, Eric Cook, had mentioned that when he does a show of hands, like, 80% of them say, "Hey, we're using it daily and weekly." But when you ask how many have read the policy or had training, we're down to 5%. Craig Kirkpatrick (28:41): Yeah. Nancy Ozawa (28:41): That's a huge amount of people. So to me, that's a- another risk that people need to become more trained on these kind of tools to be protecting themselves and the institution. Craig Kirkpatrick (28:51): Yeah. It's funny you say that because in a lot. Well, you guys know, in a lot of our trainings, we spend on the compliance side. Because let's say you're in Copilot, for example, and there's the agent store that you could access. Can you access it? Should you access it? What's the difference between work and web? Um, and so it is literally about leaning into education. Uh, how are people using it? Because by the way, a lot of people are using it on their home computers and they're bringing it into work as well. Yeah. So it, it's just more about AI training or education. How are you using it? And let's, let's use it. It's a powerful tool. Everybody's gonna figure this compliance thing out. I remember when, like, Twitter came out and Facebook and we all struggle. How do we use it? What's the compliance stuff? (29:33): We'll, we'll do that here. But along the way, it's gonna take more communication and more education. Nancy Ozawa (29:38): Right. More education on the tools so they don't make mistakes like sharing the PII, as you were saying. No, we kind of touched about risk, but I kinda wanna go a little bit more into the efficiency component. And, uh, we were reading recently that the ABA released the 2026 innovation survey. You probably have seen it already. Um, there's some really interesting stats in there. It's 75% 1% of the bankers that were surveyed said they expect significant gains in efficiency. I mean, very optimistic. They also said 43, so a little less than half are in the very early stages of exploration. So interested, but still trying to figure out how to dip their toes into it. What do you think about these adoption trends? What are you seeing when you're talking to financial professionals? Craig Kirkpatrick (30:24): Yeah. This is what we're seeing. So 2025 was a year of sort of education. People thought about what, what kind of model should I use? How should we roll it out? And 2026 is the year of implementation. So within the financial institution, we're making Copilot available to our whole organization. Oh, maybe a third of our people are gonna get access to Copilot work that can touch my emails. Uh, and again, as I said, there's a group who haven't touched it and don't know how to touch it. Uh, and there's just this really interesting opportunity to, to accelerate usage. People are saving five plus hours a week using it, and it's just gonna take some time and a commitment. One thing on that too, Nancy, you know, a lot of people think about the training of their employees. Like, oh, I don't know if people have a lot of budget to train employees, right? (31:13): They just don't. But they do have a budget to grow their business, to grow their institution. So as those group is like, no, no, no. This, this is not just training somebody how to use Word. This is trying to make everybody more efficient. And so they're sort of recalibrating, like, this is about development and business growth. You said at the start, how do you use this in sales to grow the business? Let me mention one last thing on, on the administration part and the people within these financial institutions. I'm in a month gonna be in front of 500 admin, uh, individuals talking about AI. Um, you know, people are nervous. Like, is it gonna replace me? Um, and our message to them, this is not here to replace you. This is to make you wildly more efficient. And it's not humans versus AI. It's sort of humans with AI versus humans without. (32:01): And that's what this is about. And everybody's trying to figure out how can I adopt this thing in a compliant way? And that's what we're here to talk about. Nancy Ozawa (32:10): The company really does need to change its perspective that this is not about a technology tool. This is really about how you grow the institutions. What are the most common tasks that you recommend that are the highest value that the lender should use AI to become better at? I'm trying to look for some practical advice for a lender to really start. If they're gonna start testing it out, where should they start really testing it out at? Craig Kirkpatrick (32:37): Using AI to do research, provide you, you know, uh, i- information. I have found it lowers the cost of decision-making and can accelerate it. The biggest impact I've seen in sales is just having that salesperson be better educated about their product. Hey, what are the three bullets that I should know? And what would a grumpy client ask me? Or then from the perspective of this client walking into my door, what are their questions? Or what is their perspective? Yeah. Jo Ellen McKinley (33:06): And Craig, would it not be a use of AI to say, okay, here's a list of my top five customers in this market. Give me a list of five more customers who look a lot like these people do? Craig Kirkpatrick (33:17): Yes. Here's my top five clients. By the way, AI, can you go out every week and look for local news on my top five clients and provide me with talking points I might wanna share with these folks? Or give me a simple email. So here your AI is consuming information and allowing you to connect better with a client. So it really is just almost up to your imagination on how this can, can be used. And that's what's interesting about knowing first what's possible. And then how can we leverage this tool to help me connect on a more human level, right, to, to the clients that I've been working with. And particular for the people on this call who are in local communities. Yeah. My gosh, this, this is where their value add is, right? 'Cause people want human connection. So can I be more efficient, but I kinda keep the personal touch? Nancy Ozawa (34:07): Love that human connection piece of it. Mm-hmm. Absolutely. Jo Ellen McKinley (34:10): But we're all about with community banking. Nancy Ozawa (34:12): Yeah. Absolutely. So we've been all over this topic in terms of potential use cases, how to use it, how to be better at it, how to focus on it, some of the risks on it. Let me ask you, Craig, you gave us the idea. What questions have we not asked you that we should touch on, uh, before we close this out? Jo Ellen McKinley (34:34): Act like you're ChatGPT. Exactly. But you've already learned us Nancy Ozawa (34:38): And personalized. Um, Craig Kirkpatrick (34:40): Okay. So what do you do next? What, what do you, what do you do this afternoon? What do you do tomorrow? Nancy Ozawa (34:46): Good. So let's sum it up for the Craig Kirkpatrick (34:47): Listeners. With this, with this information. And we have busy lives. So can I find 10 minutes a day to sit down and just start having a conversation? Can I incorporate this? Can I ask AI a, a problem and help me solve it? Can I make it part of my routine? Of course, here at Optimal, we train companies and so forth. And if you wanna accelerate that, we're, we're obviously available to help firms. But can you find 10 minutes today? If you haven't used this to just go in and here's what I'd say. Yeah. "I'm new to AI. I don't even know what's going on here. Can you give me three ideas on how I could begin to use you?" Uh, because I will say this, I'll, I'll end with this. The difference between someone who's using AI now and someone who isn't in the last 90 days has grown rapidly. (35:39): If you're using an AI note-taker, now all of a sudden you're saving 10, 20, 30 hours a week versus someone who isn't. So we do instruct our clients, you know, you've got probably six months to a year to sort of really engage on this training thing. And, uh, it begins at the top of communicating a mission. Let's just go learn. Let's lean in and figure out how we can leverage this. Nancy Ozawa (36:00): And, you know, I think that is very practical advice. Find 10 minutes every day. Everyone can find 10 minutes. It could be during your coffee break, you're drinking your coffee and you're talking to it. So it could be the new water cooler if you wish. Now, Joellen, you've been in AI as well, and you keep coaching your team. Is there any other examples you might share with our listeners of what they should also start? Jo Ellen McKinley (36:21): I like the example that I just gave about give me five more customers like the great ones that I've got. Um, I think that that's huge because we're all looking for referrals. So I, I love that. And it's also good to just keep track of your markets. Like, what news do I need to be, uh, kept apprised of, and what new construction projects are going on in my market that I might be able to take advantage of? Um, ear to the ground kind of thing, anticipating what's coming next. Nancy Ozawa (36:49): I love it. And, you know, you've got 10 minutes today, 10 minutes tomorrow, and now you've got a couple different things that you can do within those 10-minute little periods of time. So - (36:57): Right. Well, thank you, Craig. And thank you, Joellen. I think we've had a really good conversation and gave people lots of ideas of how to get started. Um, Craig, I attended your training earlier this year, but I think even in this conversation, I'm seeing some other tips or other value that I don't know if I picked up the very first time. So thank you for sharing that as well. To our listeners, thank you so much for joining us for this conversation on AI and sales and banking. And I hope you're walking away with some new ideas about how AI can support you as a sales professional, from either prospecting to prep, to follow-up, to coaching, getting ready for those meetings, and talking about what your customers, a grumpy customer will be before they walk in, and the ongoing relationship management. If there's one takeaway, it's this. (37:41): AI is not here to replace the human side of sales. I think, Craig, you said it. It's human with AI, not human versus AI. So AI is here to help you show up better for your customers. You can ask smarter questions, and you can focus more of your time on building those relationships. So at PCB, we're committed to helping community financial institutions understand these trends and think through how to take advantage of them in a very thoughtful and responsible way. Our goal of Banking Out Loud is to have conversations like this to give you ideas, language, and examples you can take back to your own team. To make sure you don't miss future episodes, subscribe to Banking Out Loud on your favorite podcast platform like iHeartRadio, Apple Tunes, and others. And if there's a topic you'd like us to cover, or if you'd like to join us as a guest, like Craig did, reach out to us at bankingoutloud@pcbb.com. (38:34): So thank you both Craig and Joellen, and thank you listeners. We'll see you next time. Goodbye.