How AI Frees Managers to Lead Their Teams-Not Just Manage Them

Nancy Ozawa (00:06): Welcome to Banking Out Loud. I'm Nancy Ozawa, chief marketing officer at PCB, and I'm one of the hosts of this podcast. We know that technology's on everyone's minds these days and AI is no exception. But here's the twist. AI really is anew. Many of us have been using it for years without even realizing it. Think about Netflix. It recommends what you should watch next. It's based on your behavior. That's really AI at work. What's different now is it's a new category of AI, large language models or LLMs, tools like ChatGPT, Claude, Perplexity and others. They're just not working behind the scenes this time. They're actually tools you can interact with directly. And I think that's where the opportunity gets very interesting. It's not just about the technology, it's about how the individuals and the small teams can use these tools to be more productive, more efficient, and ultimately more effective in their roles. (01:05): Because at the end of the day, we're trying to increase productivity to drive more profitability. And whether that's saving time, reducing costs, or even discovering new ways to generate revenue. The other piece to think about is your customers are starting to use AI before they even walk into the branch to talk to you or call you on the phone. They're researching options, comparing terms, and they're walking in with more informed questions. So that highlights a number of questions that we are going to delve into today. That's why we're gonna explore how LLMs can help you work smarter and create real impact within your organization. So to help me unpack this, I'm joined by my co-host, Virginia Robbins, who brings deep expertise working alongside community banks. Virginia, it's great to have you again on one of these exciting podcast topics. Virginia Robbins (01:52): Thanks, Nancy. Great to be with you. Nancy Ozawa (01:54): And we're also joined by our guest, Eric Cook, who has spent 15 years as a community banker before moving into digital strategy and now helps institutions navigate AI safely and effectively. So Eric, welcome. We're glad to have you on as well. Eric Cook (02:09): Thank you. Nancy Ozawa (02:10): So Eric, just to get us started, could you just share a little bit about your background so listeners know where you come from, especially your journey from banking into AI and go from there? Eric Cook (02:21): Yeah. Yeah. So I often refer to myself as a recovering banker. I did it for 15 years and then in 2007 made the decision that I didn't wanna be the CEO of a publicly traded bank. And, uh, I had become kind of an anomaly in the banking world. You know, I was one of the first ones to have a Palm Pilot and then, uh, a Blackberry and I had the tablets and all the crazy tools and gizmos. And we ordered a computer, uh, back in the early '90s when I first started at the bank that showed up in a Holstein branded box. I don't know if any of your listeners or the two of you happen to know what company that may have come from. Nancy Ozawa (03:06): Thinkpad? IBM? Eric Cook (03:08): Yeah. Uh, it was Gateway 2000. The Holstein box. Virginia Robbins (03:11): Gateway. Oh, yeah. With their, with their cow boxes. Yeah. Eric Cook (03:14): Their moo boxes. Yes, Virginia Robbins (03:15): Yes, yes. Eric Cook (03:16): Which were amazingly robust, strong boxes, by the way. I think there's probably gateway boxes out there to this day that are in people's garages with stuff in it. But, um, it came with a program called America Online installed. And we had a modem, we had a phone line, we had the squawky noise, and that exposed me to this thing called the Worldwide Web and built my first bank website with no board approval budget or permission. But the fact that was my father was the CEO, I probably had a little more leniency than most, but we launched marshallsavings.com in, uh, 1995, um, putting us on the web as probably one of the first community banks with a website and saw that technology was a way to bring people together and to create relationships. And we used it as an education tool, but we quickly moved into receiving emails and we were one of the first banks to lean into online mortgage lending in the state of Michigan. (04:19): So I've always had a passion for technology and innovation. Never thinking that I would go to work at the bank with my dad because that seemed like just numbers and accounting and that just was not my cup of tea, but I got to flex my muscles over on the marketing and on the technology and the operations side to really help bring that innovative spirit into the organization. And so now, since 2007, our agency has been helping banks around the country with their websites, with their digital marketing. And of course, everybody wondered what was going on with social media when it came along. People were scared of it. They couldn't control it. And, um, we're hearing a lot of those same things as it relates to AI. So I've moved back into digital therapist role and helping banks get comfortable with AI, risk it, understand how to use it strategically. (05:12): And man, it's been a fun and busy couple of years for sure. Nancy Ozawa (05:17): Well, fabulous. I love hearing that story. That's great. Uh, Virginia, you've also seen firsthand how technology can change and impact job roles. Do you wanna kinda touch on your experience, especially maybe in customer service and how that's changed with technology? Virginia Robbins (05:31): Well, so Nancy, many, many decades ago when I started in banking, I was in college. I was working in Massachusetts. I was on the campus at MIT and on payday, there'd be a slew of folks coming in to cash their paychecks. My job as a teller was to fill out the triplicate form with my number two pencil and make sure my drawer balanced. Flash forward to what we asked tellers to do today, and they're really not tellers anymore, but customer service. The computers and technology manage all that work that I used to do to find a nickel in the bottom of the drawer. That's all done now through technology. And while the job still exists, customer service and empathy, which are your number one skills or your, your most important skills. Nancy Ozawa (06:14): Yeah. Virginia Robbins (06:14): And so what we look at when we look at technology, whether it be... And Eric, I agree, you know, it's been amazing with what we can do with websites and the internet and the rest. And banks have moved forward on these things and really provided superior customer service over and over again. All of this technology allows bankers to do what they do best, which is focus in on relationships. Eric Cook (06:35): Absolutely. Hearing you say that the teller gives me the flashback. I was the teller, even though I was the quote boss's kid. I was the one that they started balancing on a Friday afternoon at maybe 2:30, 3:00 so everybody could get out of work and make it to the football game on time because it seemed like everything I did was divisible by 9:00 at the end of the day, which is an accounting joke that hopefully bankers will catch. Um, but I very much enjoyed the teller customer service conversational components. And, you know, I hadn't thought of it from a, just a pure technology, but AI or technology or automation or cash recyclers even, if that could have taken the balancing functionality off of my plate to not have to worry about counting all the pennies and dimes and nickels and I could just focus my attention on being in the moment with the customer, talking about their needs and exploring their questions and having really valuable relationship building conversations, um, that would've been a completely different experience for sure. (07:39): So thank you for that little trip down memory lane. Virginia Robbins (07:43): So Eric, when we think about this, a lot of what bankers do is in fact looking at the basics of, of the business. It's the compliance components. It's the routine processing. It's managing and monitoring. And yet when we look at AI, AI can do some of those things, at least 80% of it - Yeah. Nancy Ozawa (08:01): Much... Virginia Robbins (08:02): More effectively and faster than the human. So with that, if I'm a banker and I think about those tasks being my primary job, should I be concerned about job security? H- how should I process that? Eric Cook (08:16): Yeah. When we do training, that's usually one of the first things that we hear. We do an anonymous survey to help get just a fingerprint of the organization to help us see, you know, what's the risk tolerance, what's the attitude, what's the temperature? What's the prevalence of shadow AI? Even in organizations where we hear them say, "Well, we don't allow artificial intelligence in the office, so nobody's really using it." And then we're like, "Well, let us do an anonymous survey and see." And we get 30, 40% of the people that say, "Yeah, I use AI for my work." They're just sending it to their Gmail at home and doing their writing and then sending it back to their office. But when people think about the hesitancy, because we've had people that have said that in their actual surveys, "I don't wanna use it. I don't wanna teach it what I'm doing because then I won't have anything to do, and then I'm not gonna have a job." What we try to get people to think of is, back to that teller example, if you were to tell me that, "Hey, we're gona bring this cash recycler in, or we're gonna bring in technology, or we're gonna give you access to a chatbot that can help you answer questions faster for your customers," I wouldn't have thought that that would've put my job at risk. (09:28): It would give me more power to do the things that I enjoy and lean into the human side of me being a financial community professional. And those are the things that most people don't think of because AI and the concept of it is such a new concept. They immediately hear the stories of Amazon's laying off 5,000 people and Klarna lays off 4,000 people. That's the immediate response when they hear a business wanting to implement artificial intelligence. And so what we try to do is think about what are the jobs that are just mundane, that are data intensive, that are repetitive, that if you could show up tomorrow and not ever have to do that ever again, what would that one job be? And in most cases, the banker smiles and says, "Do I have to pick just one? 'Cause I've got, like, five in my head right now." But those are the things that we may not be able to completely get rid of. (10:26): But to your point, 80% of that could be taken off of the table, which then frees you up to be able to spend more time with your customers, to have better conversations with your team. We asked that question one time of an employee and said, "If we could give you two hours a week back, what would you do?" And she sat back in her chair. She was on Zoom and she's like, "I would love to just sit in my office, shut the lights off, and dive into the computer programs that we use that I know we're not even scratching the surface. But I'm just doing the basics now because that's all I have time to. I don't have time to learn what else we can do in these programs, or I'd love to be able to mentor my team. I don't feel like I'm giving them my best. (11:10): And there's things that I would love to teach them and be able to have them ask me questions, but we're just so busy with our hair on fire trying to get everything done that we don't have enough time for enrichment, team building." So whatever that is, there's always something else to do, and it's the human to-dos that often get pushed by the side because the compliance, the documentation, the paperwork to-dos, you'll get in trouble for that if you don't get those done. An examiner's not gonna come in and go, "Oh, guess what, Virginia? You didn't spend enough time mentoring your staff this month, so we're gonna write you up." That's not anything that comes up. But if you don't fill out your CTRs the right way, that's something that's gonna be a problem. And so we try to get people to think about the other stuff that really matters but is sometimes difficult to get to and quite honestly could sometimes hard to measure. Virginia Robbins (12:02): Eric, I appreciate that so much. It, it, when I've talked to some of our customers and employees and they think about AI, the thing that I think makes people a little bit more hesitant is the concern that, well, AI's not always right. Right. Nancy Ozawa (12:16): So Virginia Robbins (12:17): How do I turn over these key compliance or these key regulatory activities or these key monitoring activities and yet ensure that my outcomes are what they want? Because - Yeah. As you mentioned, it's, that's the thing that people will measure my success by. Yeah. If I, you know, so how, when you think about those kinds of questions and challenges, uh, what are the thoughts that come to your mind? Eric Cook (12:40): Sure. Well, present company excluded because I'm sure both you and Nancy never make a mistake and you're perfect all the time. Um, and if you were to ask my wife, I'm sure she would say the same about me. Um, probably not. But the thing to keep in mind there is AI was trained on the information that humans created. And not everything that's on the internet, not everything that's been written, not every video, audio, image accurate. And so it's trained on inaccurate data. So it's going to have those inaccuracies that are just part of human nature. So you don't have to, nor should you, ever take something that comes out of AI as the gospel. And this is one of the things that I think oftentimes overlooked because in a number of instances prior to artificial intelligence coming on the scene, when a new technology gets introduced, social media, TikTok, email, the telephone, well, however far you wanna go back, the younger people in the workforce adapted faster because they could change. (13:48): Maybe they grew up with that technology. You ask a 50-year-old like myself, and I'm being generous because I'm now in my late 50s, but if you ask me to go put together a fun TikTok that's gonna generate a lot of visibility, you're probably not gonna be very impressed with what I'm gona come up with. But if you ask my niece or my nephew to do that, boom, they've got this. So they understand the technology. But what people don't understand is on the AI side, those of us that might have a little gray hair, we have the experience. We have the perspective and the context of 10, 15, 20, 30 years in the industry. So when AI kicks out a response that says, "This is how you need to handle this loan, or here's how you need to talk to this employee," we should be very quick to put that through our sniff test and go, "Does this smell right? (14:42): Like, if I didn't have AI tell me to do this, would I approve this loan? Would I talk to an employee this way? Would I approach a problem in this particular manner? And if it doesn't feel right, you have full right and authority, and I would say even expectation to push back. And we've had that even come up before where we were talking about AI training and somebody raised their hand and said,"You know, I've been doing project management for 20 years and I have a degree in it and I don't wanna use a platform that's gonna tell me how I need to do something and I know that it's wrong." And I said,"You shouldn't. You should push back and go, You know what? Based off of my knowledge of how project management needs to go, or based off of how we do it here at our institution, these are the things that I was expecting you to say. (15:28): "And then you open the dialogue with it. You say,"Why did you recommend what you did? And where am I maybe missing something?" 'Cause you might have a learning opportunity where there could be a new process, a new protocol, maybe a different perspective that you hadn't seen because sometimes blinders get on our face and we like to do things the way we've always done them. But you should be pushing back and talking with it, communicating and collaborating alongside it. And I say it being AI, but you should be comfortable pushing back to make sure that it works. And if you don't trust it, you shouldn't use it. And the younger generation that maybe see AI as an operating system and not necessarily an assistant that we work with, they're not gona have the expertise to know whether or not that recommendation for that commercial loan is right or wrong because they don't have 15 years of commercial lending experience. (16:24): So they're just gonna say, okay, that, that makes sense. But the great part about AI is you don't have to have the technical expertise. I don't know, have to know the filters and the settings and the croppings for all the TikToks. All I have to do is press the button and talk to it and have a conversation. And whether it's typing or talking, the interface now allows for anybody to exhort their expertise into the conversation to direct it and to get the output that you want. Virginia Robbins (16:56): And I think for our listeners, things like one of the prompts that might be useful is something like, "What do you remember about me?" Or, "What do you remember about this project?" It's sometimes really helpful to understand what information the AI tools picked up because it may be that there's got one fact that's - Yeah. That's not right. Yeah. Or picked it up from Reddit or someplace that wasn't, you know, a valid source, but something that might be a little different than your experience has led you to believe. Yeah. And so validating those assumptions can fine-tune that output. It's very interesting. One can have a dialogue with AI that really does feel like we're talking to someone who is a very intelligent seven-year-old maybe. Mm-hmm. Um, and I'm sure, Eric, you've had the experience where you've prompted a, uh, an observation or a fact to have AI tell you, "Oh, yes, that's correct. (17:56): I'm wrong. Um, let me go check my sources." Yeah, Eric Cook (17:59): Absolutely. So Virginia Robbins (18:00): I think that challenging that you mentioned is really, really important. And I loved what you said. If it doesn't feel right, doesn't match policy, doesn't match guidelines, challenge it and understand where it's getting its information from. And maybe then fine-tune your prompts and assumptions to help that ensure that that doesn't happen in the future.... Eric Cook (18:22): Yeah, absolutely. The story that you just sparked, because you mentioned Reddit, and it's a perfect example, is in the early days, Google's artificial intelligence platform struck a deal with Reddit. And that still is a very popular source of information for AI searches because it's 100% human generated, unfiltered. But there was a subReddit out there that was completely satirical about ways that you could keep cheese on a pizza and glue was the answer. So there was all these just funny, like, "Oh, you gotta put this type of glue, and I like super glue, and I like Elmer's glue because I used to eat it when I was a kid in elementary school and it adds an extra special taste to the pizza." And Google ingested all that information. And if for a period of time, if you were to go to Google's AI and say, "What's the best way to keep my cheese from sliding off of my pizza?" It would come back with a very logical and well-sounding, "Well, it's best to use glue, but be sure because you're going to be eating it, that it needs to be non-toxic and preferably tasteless so it doesn't detract from the pizza sauce." If you had no clue that glue didn't belong on your pizza, and I'm not even gonna open the can of worms of whether pineapple goes or doesn't go with pizza. (19:40): Oh. Yeah, let's get that one there. I don't wanna practice. But if you were to read that and not understand, and this is a blatantly obvious one, so you don't need 15 years as a commercial lender to understand this one. But if you didn't understand that glue absolutely in no way, shape, or form should be on your pizza, it read really, really well. And that's an extreme example, but that's one of the things that can happen. Now, deep research and kind of recursive learning and other sorts of protocols that are being baked in are trying to help reduce the chances of hallucinations and fake information, but you always need to trust your gut. Just because it's AI doesn't mean you should trust it. And that goes for images, that goes for video, but you just need to be very careful of anything that you see, hear or listen to online because AI is so good now. (20:38): Your gut is really the only thing that you have that protects you to determine, is this real? How do I verify it? What's the source that can confirm that this indeed is accurate and not made up or synthetic? So... Virginia Robbins (20:51): Yeah. You know, we've been talking about some of the challenges that folks have when they come in to use AI the first time. Another challenge that comes up great quite a bit is data privacy concerns. Nancy Ozawa (21:02): Yes. Virginia Robbins (21:02): Um, my bank's got confidential information or my institution has confidential information. How do I balance using that confidential information and an AI tool? Are there any best practices that are out there now? Eric Cook (21:18): Sure. So the first thing that I would say, and I hear this a lot as a reason why organizations don't get into AI, is they say, "Well, we're really not sure we wanna do AI because we're worried about PII. We don't wanna put customer information into the system. There are so many opportunities and ways that you can use AI as a community financial institution that don't touch anything PII-related. To have that be the reason why you don't get on the playing field, you are missing out on tremendous opportunities. Educational content, blog content, strategizing, thought leadership, ideation, just having a sounding board. So many different ways that you can use AI that doesn't involve confidential sensitive info. That gets you comfortable with it using it as a, as a thought partner. You maintain thought leadership. That comes from a book by the name of the AI Driven Leader by Jeff Woods, which I always recommend as a great book. (22:19): But using it as a thought partner, when you get to the point where you then are comfortable with it and you decide you wanna use customer information, that's the point where you need to look at the platforms and the processes and the systems that you have in place. There's technology out now that can deploy inside of your Azure ecosystem, physical device in your actual office, different ways that you can license ChatGPT enterprise, cloud enterprise, Copilot work, to be able to have comfort. I mean, for Microsoft shops, which is probably most of the folks listening to this, you are already storing your confidential sensitive data in Excel sheets and PowerPoint files and Word documents and loan write-ups, scanned loan information inside of a SharePoint or a OneDrive environment. It's already part of the cloud. So if you can use a technology that stays within that ecosystem and understand how you only allow that, it's the organization's responsibility to teach the employees. (23:23): Because the flip side of that, and I don't know if you remember, but if we get in our time machine when we were introducing things like online banking, when we were talking about entering usernames and passwords, what's the thing that we told people to look for in the URL address bar to make sure that we knew that that was a secure website? Either of you know the answer to that?... Virginia Robbins (23:47): It was that little padlock. Eric Cook (23:48): The little padlock, which came with HTTP - Nancy Ozawa (23:52): TPS. Eric Cook (23:53): Exactly. Right. So if that's your message to your employees and you say HTTPS is secure, you can put a username and a password and you don't follow up that training. They're gonna go to HTTPS:/ChatGPT or Grok or Perplexity or Gemini, any of those. They all start with HTTPS. So they're gona say, "Oh, IT Director Eric told me HTTPS means it's secure. Secure means I can put in sensitive information like usernames and passwords. So I'm gonna upload my customer list to Grok and I'm gonna have it analyze my customers and see which ones are having a birthday or which ones are got a high balance or which ones are, you know, that's not the same. But if you don't explain that to your employees, you can't fault them for not knowing that. They're not network administrators. They're not cybersecurity professionals. They are doing what they were told. (24:55): HTTPS means secure and that might be the last time they've ever heard anything from you about what secure and insecure means. And I give that example a lot of times when I'm at conferences and I see people in the audience go, "Oh, snap." Like, I hadn't even thought of that. Like note to self, that's the first thing to do when I get back to the office is just send an email to people to say,"This does not mean that it's secure for you to put our customer information into these platforms. That's the first step. Because at the end of the day, employees don't wanna do a bad thing. Employees don't wanna put our valuable customer or member information into a platform and cause a problem, but they could do it accidentally and they'll feel terrible about it. But then, you know, the cat's out of the bag. Virginia Robbins (25:49): So most financial institutions have an AI policy. And as a manager inside an institution, I need to know what's in that AI policy because that will clearly direct me as to where the boundaries of that discussion are. However, in some institutions, I may read that AI policy and I find that it quite frankly doesn't allow me to try some of these tools that I think might really help me. So Eric, when you've come across those experiences, do you have any words of advice for those managers so that they can help - Yes. Influence their company and their strategy? Eric Cook (26:24): Yeah. Yeah. And actually, to back that question up, when I do live sessions, I ask three questions to get started. First question is, how often are you using AI? So never. And I still get a few hands that come up. Um, jokingly, I said, "You probably got a flip phone too," and everybody kind of giggles. But most of the time, people are raising their hand when we get to the weekly or daily. Like, I'll have probably close to 80% of the hands up in the room when I say weekly or daily. And then my category is you're gonna pry it from my cold dead hands because I use it pretty much every minute of every day because it's my thought partner that lives alongside of me. The next question that I ask is what type of technology are you using? But the last question that I ask is, okay, how many of you have seen your bank's AI policy and you've received any training on how to use AI? (27:16): And I see probably 5% of the room raise their hands, but yet 80% are using it on a weekly or a daily basis. That's a problem. So that immediately, that's kinda right up there with the HTTPS doesn't mean you can put customer information in the system. But one of the things that I suggest is kind of going along the lines of a statement that even goes out to your staff to say,"We know AI is here and it's not going anywhere. Whether you like it or not, AI is here to stay. We gotta get familiar with it. We gotta get comfortable with it. At XYZ organization, XYZ bank, we wanna make sure that you play with it. We wanna encourage you to use it. We want you to download the apps. But when you do, don't put PII, confidential, sensitive, non-public information into it. If you don't expect to see that on the homepage of our website or if it's not a conversation you would have with your mom in a really busy grocery store, that does not belong in AI. (28:21): "... But as you use it, as you experiment, plan a vacation, look at recipes, ask it a question about trivia, any of those sorts of things to get comfortable with what has the capabilities of doing. Then when you start connecting the dots, so then you can start building the use cases to then determine what makes the most sense to be able to deploy a strategy to use AI inside of the organization. But the important part about that is the policy needs to give the guardrails, but I think policies need to be more empowering as to here's what you can do in order to learn and experiment and to be able to get familiar with it. And when you get to the point where you've reached that capacity, here's where you go next. Because almost every policy's written with, "Don't do this, don't do this, don't do this, don't do this. (29:12): "And then you're like, "Oh, man. I wanna do some cool stuff." And the policy is a little bit of a downer. So encourage them to use it, encourage them to play with it, but make sure that they know where the boundaries are so that they feel comfortable experimenting to have those light bulb aha moments that they can bring back to the institution. Virginia Robbins (29:32): I love your response here because it really is about engaging people in the new tools so they can see the results and then from the results then drive forward. Eric Cook (29:41): Absolutely. Virginia Robbins (29:42): One of the questions that comes up quite a bit is should we make AI use mandatory? And an interesting conversation on this has been we've gotten some feedback that there may be some religious or other concerns. And so we've taken the position and are suggesting often to others that it remained optional for the time being. How do you feel about that in terms of driving AI usage in, say, a department or team, especially with some employees who might be a little bit concerned about newer technologies? Eric Cook (30:13): Yep. And that happens. You know, we already talked about how much people love change and learning new things. You know, I didn't get hired to be an AI nerd. I do bookkeeping and I like numbers and I just wanna do that. I think you need to be careful about making it mandatory, but I think there's also going to be an increasing expectation of just the evolution of business and of life. Imagine if you had somebody that, you know, "Well, I really don't like email. I preferred phone call and I like the fax machine." That might work for a little while, but if somebody didn't wanna use email, it's gona be pretty hard to integrate them into the communication cycle of the organization because it's not just external email, but it's internal email, it's delivering of information, it's meeting notices. Email has become a critical business function, like it or not. (31:06): Um, the good old days when you would turn your computer on and if AOL said, "You got mail, you were excited." Now it never shuts up, but I digress. So I think you need to be really careful about, okay, well, you don't have to be on it now, but did you have to use the internet? Did you have to use Google? There is likely going to be a point in time where it's just gonna be a, an expectation. Like you need to know Microsoft Word and Excel and PowerPoint. You need to how to send an email. You gotta know how to, you know, pick up the phone and retrieve a voicemail and snap a picture and send a text if you're out in an event so your marketing department can post that onto social media. There are things that are gona come up. And what's gonna happen and the analogy that we often use, and I've seen it in publications across all industries, not just banking, but imagine you've got a department of five people and you've implemented AI and you want the department to be productive and four of those five people have leaned in, they're saving time, they're 30, 40% more productive, they're churning out more loans, they're getting back to customers faster, they're producing more volume. (32:17): And then you got someone sitting in the corner that doesn't wanna use AI. Their volume's not there, their production's not there, they're the bottleneck and everything that happens. It's not that they're not using AI, it's the fact that they're just not keeping up. I don't wanna drive a car. Well, guess what? You're gonna lose the Indy 500 if you show up on a horse. That's just the way it's gonna be. And so it's, it's that trying to figure out, okay, well, what is it that you're concerned about? What are you worried about? What are the things that intimidate you? You can just talk to it. Voice dictation. You don't need to be a good typist. Those are things maybe more out of ignorance and just not knowing as opposed to the full-blown fear and resistance of, you know, Skynet's gonna take my job. Um, so that conversation is one that I think is gona be had in a lot of organizations as this becomes more and more pervasive just throughout society in general. Virginia Robbins (33:17): You know, Eric, you mentioned it being more pervasive. There's another side to us as bankers using these tools. Our customers are using them. Eric Cook (33:26): Oh, yeah.... Virginia Robbins (33:27): Our customers are using them to review agreements. Our customers are using them to prepare for discussions with us. Nancy Ozawa (33:33): Yep. Virginia Robbins (33:34): So if I'm that one out of four that's not using it, I now am coming to that customer discussion possibly not as prepared as I could be. So what have you heard about customers and their use of AI? Eric Cook (33:49): Well, you probably remember the first time when you went car shopping and you could go on the internet and you could find the invoice and the holdback and the reserve and how strong and empowered you felt where you're walking into that car dealership and you're like, "I'm not gonna pay a penny over this because I know exactly how much money you're making and da, da, da." That poor car salesman didn't stand a chance. Their margins, their negotiating power just went right into the toilet unless they figured out a way to change their perspective and sell on value and relationship and not just price. And that's going to be something we're going to have to be comfortable with. One, customers are gonna know the competitive landscape way more than what they've ever been able to do because I can do a deep research and go, "I'm interested in buying a house. (34:37): I live in the Grand Rapids, Michigan area. I'm interested in comparing loan rates across banks and credit unions. I wanna stay with a small local lender, but I'm not opposed to using a mortgage broker. What are my options? Oh, and by the way, go out to Google and Reddit and Yelp and find reviews so I can see what other people are saying." And I walk away from the computer, I go get a cup of coffee, and in 15 minutes, that research report's done. Mm-hmm. And now I go meet the lender and I know about them and everybody else in the market that's doing this. As a lender, as a banker, if you don't know customers are doing that more and more, and if you're not using that in your daily work, when you go meet with a prospect, do your research. I'm meeting with a manufacturing company or I'm meeting with a veterinary clinic. (35:27): I'm a commercial lender. I don't know anything about veterinary clinics, but I wanna talk intelligently. Flip the script. Go in and, you know, you now know what's going on in the industry, medical trends, insurance situations, breakthroughs in veterinary technology. You can actually, in very short order, talk the talk to be able to build that relationship and try to take price not off the table, but it's not the only thing that's on the table. And so we have to understand that the world is becoming more intelligent and we have to be on that bus. Or you know what the option is, is you're under the bus and nobody wants to be under the bus. Virginia Robbins (36:08): Yeah, nobody wants to be under the bus. So we've seen this increase in information just continue to improve. We've seen that information just continue to grow and our access to information. Yes. And yet, if we're not careful, we all sound the same. So when you're looking at AI and you're having a discussion with it - Nancy Ozawa (36:27): Yep. Virginia Robbins (36:28): What are the best ways to bring your personality into that discussion so that when you go and have that discussion with the veterinary, or when you go and have that discussion with the prospect, that it's actually you and not just AI that's, that's leading that discussion? Eric Cook (36:44): Yeah. Um - Virginia Robbins (36:45): Have you find the best ways to insert your personality? Eric Cook (36:47): Yep. And that's one of the other pushbacks that I often hear from, well, we can't use AI because we don't wanna sound like a robot. You know, we're a community institution that's been here for 100 years. We have a personality. We have a reputation. We have knowledge of the local market area. AI doesn't know that. Well, it might know a little bit, but it's not gonna know as much as you. And so taking the time to train the AI with that information, you know, we've all been writing emails for how long? Maybe some of you have been blogging or writing articles or you share content on LinkedIn or other places. Take the time to educate the AI about you. And in many of the platforms, you've got the ability to go into your profile. You can give it information. And my profile says, you know, I'm a recovering banker of 15 years. (37:38): I like to refer to myself that way because it generates humor and laughter. Uh, I'm passionate about what community institutions do. I run a digital agency. I love doing da-da-da-da-da-da-da-da-da. Oh, and by the way, I'm also a proud dog dad to a couple of amazing golden retrievers. I love mountain biking and Metallica. And so it has that personality about me. So when I write a blog post, you know, there might be a heavy metal reference or a reference to, you know, something dog related or those are things that the AI's not gonna know unless we take the time to train it. And most people, one, don't even know that that happens. But once you know that that happens, they don't do it because it takes work. AI is not an easy button to be able to go press button, create blog post, because you're gonna press a button and you're gonna get a blog post that's gona sound like everybody else. (38:31): Getting a mortgage loan here at XYZ, blah, blah, blah is great. Blah, blah, blah. We're friendly, blah, blah, blah. We're responsive, blah, blah, blah. Doesn't have any of your organization's personality to it as opposed to as an organization founded in 18 whatever, we've got long roots and heritage mentioning things that have happened in that community that you've grown up over time. Now it's like this was written by an institution that's legitimately been here for over a hundred years as opposed to it's just AI slop about how to get a mortgage that I could get anywhere. And guess what? That stuff's not ranking. That stuff's not getting indexed. That stuff is being ignored by Google. You are just gonna become more and more invisible to people that want what it is that you have to offer. It's gotta be relevant and unique to you and you've gotta take the time and put in the work to make that happen. (39:26): No getting around it. It's not an easy button. Nancy Ozawa (39:30): I just was laughing because I, I went and looked at what it though about me and it picked up all kinds of facts that are not particularly true. I though that was very interesting. Um, we've been covering a lot of different areas here. If I'm one of those middle managers listening to this podcast, I've got so many different ideas going on in my head right now. But Eric, we should kind of sum up this episode for them. What are one or two practical things they should be thinking about today to start implementing this week? Whether it's with themselves or with their team or with their peers? Where do you think they should start today with just a concrete example? Eric Cook (40:06): Yeah. I get that question a lot and I always struggle because not surprising. I, I'm the type of guy that wants to do everything right now. Let's boil the ocean and take over the world kind of thing. And that's not realistic. Yeah. I think (40:20): Sitting back and one of the slides that I will include in a lot of my presentations is you can't do everything. You have to first sit back and go, "Okay, what are the business challenges that I am legitimately facing on a day in and day out basis? What is preventing you from being more human at your job? If I were to say, okay, for the next week from three to five, you're gonna be out of the office talking to customers and, you know, prospects and networking and you can't do any work at the office. Where would you start stressing out and having hives? Well, I wouldn't be able to get this done. I wouldn't be able to get this done. I got this report. I got this report. Think about if you were to pull back and be more human today, what are the things that are gonna have to give? (41:14): And that's where you need to start thinking and having some dialogue and conversation maybe with others in your organization, maybe as part of your AI council or a work group or a support group, if you even wanna call it that. But have those discussions around what are the things that are pulling us back from being more human with our clients? And you'll start identifying commonalities. You know, the, one of the grids that we'll often give is if you draw a line down the middle vertically and the middle horizontally, you get four quadrants. And the quadrants are if it is, uh, data intensive, if it's time-consuming, if it's generative or it's repetitive. And over the next two weeks, anytime you do something, jot it in one or more of those quadrants. You gotta review the NSFs. Well, that involves data and it's repetitive because I gotta do it every single morning. (42:11): Anything that you do, put it in that grid. And after two weeks, you'll have a very clear, okay, wow, I do that a lot. Now you might start to identify some of those use cases. Nancy Ozawa (42:25): Love it. Eric Cook (42:25): And the other thing that I would suggest is you're not using one of the language models, get one, download it on your phone. Again, be very careful. No PII. But all of the language models now have a little black circle with a little line waves that looks like an audio file. It gives you the ability to actually talk to it. It's a voice. Just talk to it and have a conversation and say, "Hey, um, I'm a banking professional just kind of dipping my toe in AI. Um, I need an AI friend to help me figure this out. What are some of the questions that I should be thinking about that maybe will help me frame up where AI might be able to help me?" Nancy Ozawa (43:06): Absolutely. Eric Cook (43:06): "Oh, that's a cool idea, Eric. Let's go ahead and see. Well, what is it you typically do on a regular basis? And you don't have to talk about your customers, but now you're entering into a dialogue and a discussion and the paid models will get you a little bit longer conversation, but then that gets you familiar with the new interface. You know, there's a lot of people that say the keyboard and the screen is probably not going to be the interface that we're gonna be using moving forward. It's gonna be voice. I've got the meta glasses. We can just talk to it. It's got vision. So the interface of being able to just talk to it and have a conversation, the better communicator you are and the better command you have of the English language, the more you're gona get out of this platform. Right. (43:52): And that is a really fun experiment that will help reveal some things that will also point you in the right direction. Virginia Robbins (43:58): I wanna get back to one thing Eric said earlier. Yeah. Um, one thing managers can do immediately is actually go find their AI policy and, and, and read it. And read it. And read it. And find out what I can do and what I can't do. And then start to work with their compliance managers while they're educating themselves - Yeah. While they're picking out the work to do. Because if the organization is allowing these things, fabulous. Yeah. If the organization is still reviewing these and has concerns, step forward and volunteer to be part of pilot groups or to be engaged in that discussion. Knowing where your company is, um, can really help you move forward. Nancy Ozawa (44:38): Yeah. Eric Cook (44:38): One last thing. Don't just assume that AI is going to be done by the IT department and you don't need to worry about it. In fact, we're actually, and it's not that IT doesn't need to be involved, but we've seen and others in the industry and outside of the industry have, have found that when you get marketing, human resource, operations, more of the, the human side of the business involved in the AI, that's when you're gona have a lot greater chance of success. Because at the end of the day, it is much more of a leadership change management culture issue than it is a technology issue. So if you're listening to this and you're a branch manager, you're someone that's in HR and you're thinking, "Man, this would be really cool to do," this is going to create some opportunities for people in the bank and other institutions out there that you would not necessarily think would be, but you could be that AI champion. (45:40): You could be the one to lead the charge to help get the people excited, to look for the use cases, to liaise with technology, to liaise with compliance, but someone that understands the human element of change and strategy and can tie it all together, that is, I think, going to create some really exciting opportunities for banking professionals down the road. And if that feels like something you'd like to do, raise your hand and keep raising it because we all need those people. And that's, I think, gonna be a really important role moving forward for anyone. Nancy Ozawa (46:15): You've hit so many different pieces on that, and I think those are very concrete steps for somebody to take. And I love the enthusiasm and the way that you're positioning us. People have been thinking about AI as something scary, it's gonna take over, but really repositioning it as this could change your role to be more strategic, allow you to - Absolutely. Have those two hours to dig into the computer, as you were saying, and learn more because that's gona benefit the institution as well. So I think that, that just changing our perspective of how we look at the issue is very important. So -... Eric Cook (46:47): Yes. Yes. Nancy Ozawa (46:47): Wow. We, we, we covered a lot, Eric, and you're right. We could con- continue talking for the rest of the day. Eric Cook (46:53): Where did the time go? But I really appreciate the opportunity to chat with you both. This has been very lovely. Nancy Ozawa (46:58): Yeah. Thank you, Eric. I've been thinking about the example of talking to my peers about the S and HTTPS or even the example about the glue on the pizza. Uh, I'm gonna have to go check out with my AI agent if it does that. And for Eric Cook (47:11): The record, I do like pineapple, so I'm sorry if there's Nancy Ozawa (47:14): Any idea. I'm good with you on that, Virginia. I'm a Eric Cook (47:15): Pineapple guy. Nancy Ozawa (47:16): Yeah. Well, thank you so much. And thank you for sharing all of this wonderful content with us. Eric Cook (47:22): Absolutely. It's been my pleasure. I'd be happy to come back anytime. Nancy Ozawa (47:25): Perfect. Thank you. Well, thank you, Eric. And, uh, thank you, Virginia. I think we've learned several interesting nuggets. Um, I've definitely taken a lot from this conversation. I've got ideas of how to talk to my team and other teams about how do you harnes AI to help get those two hours back to do more of the things that will help you do your job better by getting rid of some of those repetitive tasks and delegating them to AI. And how many of us are looking at HTTPS and thinking S still means secure, therefore I can put anything I want into that? And I still have to laugh about the idea about glue and pizza. We definitely need to challenge our AI agent to ask it. Why did it come back with that answer? Why didn't it come back with a different answer? So have conversations and challenge your AI agent as well. (48:12): Listeners, I hope you're walking away with a clearer sense of how some of these tools like, uh, ChatGPT, Claude, Perplexi, whatever you have accessible can help you and your team work more productively and more efficiently and how that can really translate into real impact for your organization. Just as importantly, I though it was very interesting that we touched on how customers are beginning to use AI to prepare for their interactions with you. And likewise, you can prepare to have conversations with them using AI. We do have another AI episode coming up that's gonna be discussing how to harness AI specifically for lenders and others within the bank that promote your institution's products and services. So in order to know when that episode is released, make sure you subscribe to our podcast and get the email notification when it goes live or follow us on your favorite podcast channel, iHeartRadio, Apple Tunes, whatever you listen to. (49:07): And if there's a topic you'd like us to cover, or if you're interested in joining us as a guest, we'd love to hear from you. You can reach us at bankingoutloud@pcbb.com. Thanks again for listening. Thank you, Eric. Thank you, Virginia. Until next time, have a great day.