How Credit Union Boards Can Use AI to Strengthen Governance and Move Faster With Clear Guardrails

Doug English: [00:00:00] Welcome back to C.U. On the Show. I am delighted today to have Deedee Myers and JD Myers. Deedee, welcome back. Uh, delighted to see you again. Love the content that you help us in the governance areas. Is that what you’re gonna be talking about today?

Uh,

Deedee Myers: we’re gonna be talking about the AI and human-centric part of organizations in the future, Doug, and the the future is here. Now, how do we wanna look at ai, look at people in the future and what the challenges are that I see an organizations need to face into right away. Really exciting topic.

Doug English: This is a podcast from the future.

Dee Myers is here to help us with the future. And JD Myers, welcome, sir. The first time we’ve had you on the podcast, tell me a little bit. First thing I always love to hear is how did you get started working with credit unions? And what kind of work are you doing with credit unions today?

JD Myers: Well, thank you, Doug.

Pleasure to meet you and pleasure to be here. Thank you for the invitation. I am a thought generator when it comes to AI [00:01:00] leadership, organizational development, and organizational design. I help all kinds of organizations be a nonprofit or profit, find their way into the future as DDR. Pointed out, uh, incrementally and strategically and, uh, with the culture and environmental mindset in mind.

And I also promote CMMC, which is the cybersecurity maturity model with DOD contractors, which they’re 84,000 of ’em out there that need a lot of help. So I bring AI into the picture there. Leadership development on C-M-M-C-D-O-D, security and development, uh, within of the CMMC framework into the organization.

But, but we’re facing a lot of complications there at a lot of time constraints and necessity to get that program moving. So we’re introducing AI into that picture as quickly as possible.

Doug English: It’s moving incredibly fast and everyone’s struggling to keep up. So hopefully what our listeners will get from the ideas you bring today [00:02:00] is some framework to how to start and how to build a plan.

Uh, uh, one of the takeaways I got from the pre-reading is how to build a plan to get your board, uh, up to speed on being able to understand things. So, uh, DEI let’s, uh, let’s see that deck that you brought and start, uh, into the, uh, content.

Deedee Myers: Yeah, just a little bit piece here to, to help us, uh, frame what’s going

Doug English: on.

Hold on, DEI, right now we’re, we’re seeing the zoom window. You gotta, uh, to other window. Okay. We will stop sharing that. Thank you. It’s okay. All good. Editor will take care of removing that, uh, show. No problem. Awesome. Awesome.

Deedee Myers: Okay.

Doug English: There you are.

Deedee Myers: All right. So a little, there are all the initials after J D’s name CMMC, so quite impressive, uh, background there.

Uh, but the piece that I think we need to look at here, JD and Doug, is what’s going on with the world right now. So this is a study that came out in 2023 for the Organization for Economic Cooperation and Development. So it is a, [00:03:00] it is a, a, a independent study. It’s not ours, but we studied the data and we, we looked at where are we in the world in terms of our workforce, in terms of different areas of literacy, uh, general literacy, numeracy, literacy, and then what we call adaptive problem solving literacy.

So the study looked at a number of different countries and you can see where we are in the United States here amongst these other countries. So right away you’re gonna notice that we are at, you know, to the left of that middle line of the average. We are on the bottom here in terms of different areas of literacy.

I think it’s important for us to also understand that literacy is actually measured at the fifth or sixth grade level, the fifth or sixth grade level. So the deal here is these are the people that we’re gonna be hiring in the future. And the piece here is, what does that mean in terms of ai? What does that mean in terms of governance?

What does that mean in terms of how we recruit people, develop people, and, and keep moving into the future so we can be [00:04:00] sustainable? So the first, first part here is literacy, really is, um, about how do we look at different documents and make sense of it. You can see that we’re, we’re below the average there.

Numeracy is pretty simple. Uh, an example is, can we put together a spreadsheet with a calculation to calculate. Uh, percentage of a loan. Like what is, what is our interest rate, uh, risk, or what is our interest rate on a loan that we have as an individual or in our organization for our member? And then adaptive problem solving is, you know, can we actually solve problems?

So the challenge here is that 32% of our population in the United State is below the baseline of proficiency level, again, which is fifth and sixth grade. I’m thinking about, um, all of my kids and I’m going, oh my gosh, really? How do we keep growing, growing this, this baseline and moving it up, uh, so we can be better at adaptive problem solving?

The other challenge that we’re seeing is that we’re clustering the United States is clustering more at the bottom of the graph, [00:05:00] Doug and jd, which means that population is increasing. So yes, we can go back to the root causes, what’s going on in our homes, what’s going on in our education system, but at the end of the day, we’re bringing these people into our organization in different roles.

The good news is, uh, that as those who, uh, score well in adaptive problem solving, actual report, higher levels of life satisfaction, because they, they know what choices are and they can make choices, and they have healthier lives, you know, financially, mind and body and, and soul. So a lot of good data here.

So I’ll pause here and see what you all think about this.

Doug English: Hmm. Yeah, it looks like the solutions are, you could hire internationally, right? That would be a opportunity. Remote employees and Finland and Japan appear to be very effective. Uh, but, uh, uh, kidding that the, that’s, uh, that’s a problem. And I imagine that AI could be making it even worse.

JD Myers: Well, [00:06:00] that’s a good point, Doug, if I may jump in. AI does two things. Minimally, it accelerates and it amplifies. If you have a workforce problem, it’s going to amplify that. Mm-hmm. You’re gonna see more of the workforce issues come to life revealing AI will reveal workforce issues that you didn’t, you weren’t aware of.

And it can also accelerate a lot of the problems because of the workforce shortages. So as you see here, the literacy rate, numeracy rate, and so on and so forth, is low AI needs articulated prompts. It needs somebody that can articulate a clear, precise, objective prompt. And a prompt is a, a question asking for data.

Take that data and create knowledge out of it. Take that knowledge and create action on it. So this is a huge issue for ai. Now AI will eventually learn [00:07:00] the user’s deficiencies and try to overcome them, but it has to learn them, has to learn the deficiencies, and has to learn how to fill those gaps with the right words, the right verbiage, the right question.

So it may reshape questions. You may see that from time to time. As an example, in chat GPT, you ask a poor laid out question, it’ll reformat and reformulate the question, come back and said, is this what you want? But it takes time to do that. So in the meantime, this leads to a new phenomenon within AI and a new buzzword called, called work swap or AI swap.

And it’s really not AI swap it’s work deficiencies that are being amplified by AI’s. Work on the deficiencies.

Doug English: Hmm. So what, from from the credit union standpoint, what do you, what do you do about this?

JD Myers: Well, we have to improve literacy across the workforce. That has to be one of the [00:08:00] objectives. Mm-hmm. One of the strategic grow ar growth areas, learning and development.

A lot of organizations have LD departments in sub-organization, or they outsource it. They go through their hr. So as part of the LD program learning and development, we have to start integrating what the schools were unable to promote and, and, and gain within the students. We have to reeducate the students into a literacy form within the workforce and with ai.

Doug English: Now I, I sound a lot bigger than just credit unions, right? That, that’s, that’s the whole educational system of the country. Do, do we, do we have like a specific things to the, for the movement? Like, uh, as an industry, what, what are we to do? We can’t, we can’t make the educational system change, at least, uh, not, not for a while.

Maybe eventually we could get it to adapt, but, uh, what do we do for now?

JD Myers: That’s a very good question, and one can come up with a lot of solutions, but I think one of the solutions is [00:09:00] we look for an area within the AI frameworks, which there are many to plug in literacy development. So we, we can ask ai, which is there anyway in front of the user as a student and a teacher relationship, to take time to help the, the workforce user learn the right way to ask questions with ai, the right thinking approach.

What is the end game in mind in this conversation with AI today on this particular topic? So AI is there. Might as well turn it, turn it into, so ask the AI

Doug English: to to teach us how to best use ai. Ai. Correct.

Deedee Myers: Well, I’m wondering too, a couple things in listening here is what does continuous process improvement mean in, in the credit union land where there’s a lot of that work going on around in the lending system? Would that be a place for us to start it, to improve the processes there and then integrate AI in there?

The other thing I’m [00:10:00] listening to is how do we onboard our credit union employees and, and keep their learning and development going? And the third thing is, is do we have the talent in our credit unions to actually teach and develop people? If not, what do we do about that? So what, what, what do y’all think about that?

Start in the lending area, would that be the place most credit unions would wanna look in terms of their strategy and their strategic projection? And then jd, what do you think about continuous process improvement integrating with AI in the lending area? I.

JD Myers: Well, if AI is producing the wrong results and a, and a financial institution, that financial institution will come under greater scrutiny by the federal government because there’ll be more mistakes made, more amplified mistakes, accelerated mistakes that are gonna hard to catch up with and, and to stop and clean up and fix.

So AI needs to be an integrated teacher throughout the day [00:11:00] for each and every student that it detects a deficiency in their ability to articulate what they want. So that will be an area where models and modules can be introduced into an AI framework of any type to help, uh, each user learn how to better work with ai.

Deedee Myers: Right? So in the past we’ve talked about point and then system and enterprise, uh, components of ai, if I’m using the right language. So if we use Point, can you give an example, point N in a financial institution so we can all get organized on that.

JD Myers: Well, a point in the financial institution can simply be, for instance, copilot, which is integrated within the browser of Microsoft and copilot is becoming an AI engine.

All of itself in the sense of it can take what you’re looking for and wanting to do, go out to the, what’s called the large language model, get the information you’re [00:12:00] after, you’re seeking, and bring it back into an intelligent proposal of this is what you might need to know in order to make the next step profitable to add value.

So that’s copilot is on a system, it’s on a laptop, it’s on a phone, it’s on a computer. So at that point, within the knowledge network, somebody is asking copilot to give it assistance. And then if you have a hundred employees across the institution and each and every one of those a hundred employees is using copilot, each of those a hundred employees are a point.

Within the ai, the greater AI system.

Deedee Myers: Okay. All right. So Doug, uh, thinking here what you said about credit union. So what I’m hearing here is that we could use point AI for better fraud detection, right? Maybe automated member service via chat box, um, improving credit decisioning. And then what, uh, what, what do we need to do in terms of getting [00:13:00] our systems up to place and our people, uh, leveraging, um, all of this intelligence in certain different ways.

JD Myers: Oh, is that a question for me or Doug? It

Deedee Myers: was just a, um, uh, dialogue, just, uh, asking, you know, make it relevant. Yeah. I, I’m

Doug English: very interested to see, you know, I, I I, I always sort of start the thinking from the top, right, from the strategy standpoint, and that’s the board, uh, and, and the, and the senior executive team.

I wonder about, uh, AI readiness, uh, AI involvement, like, you know, uh, uh, J D’s, uh, idea that you ask the AI how to use it, uh, best how to prompt it best. Uh, well, ha have you seen, uh, a credit union board actually having an AI as a part of the board meetings and, and, uh, you know, you can put it into a mode where it’ll listen and talk back and forth.

Is any, uh, have you seen that yet, Deedee? Yeah, that’s starting. It’s

Deedee Myers: very exciting. We do no one credit union where every board member got AI [00:14:00] certified, Doug. And what they’re doing is, uh, using data-driven, AI enabled conversations, you know, to, uh, look at how to create personalized and intelligent information.

So, so I asked the chair, why did you go get certified? He goes, well, our CEO and our team is bringing, you know, uh, strategically growth, um, insight powered through ai. So we need to be up here in the conversation so we can meet them here and then go further, further, uh, from a strategic perspective versus always being carried and broad and led Yeah.

Dynamic of the conversation, the quality of the conversation, and the pace of change. So it’s pretty exciting.

Doug English: Yeah. Do you know, a, a for our listeners, any specifics on where they, uh, happen to get their, uh, AI certification?

Deedee Myers: Yeah, there’s a dozens of ’em out there. I think Cornell has ’em, a lot of big universities.

Harvard, [00:15:00] um, you’ve got, I think Gartner has ’em, there’s, there’s several I could, uh, send you could add at the end here. Absolutely. Yeah.

Doug English: Yeah. It makes sense that if you’re, if you’re, if, if your organization is gonna be AI forward, that should start with the, uh, the strategy coming from the board and looking for, uh, how, you know, what is the competency you need to have in order to, uh, build strategy around the AI initiative.

Deedee Myers: Yeah. We’re also, when we’re doing, uh, board succession, uh, plans, uh, AI is amongst the top three requirements or hope for competencies for new board members. Hmm. I bet you it’s hard to come by. It’s hard to come by. Where, where we get a little stuck though, I think, is that sometimes we think, or the board thinks that they have to have somebody who actually uses AI every day.

Well, yes, that’s great, but what, what we’re looking for are people to serve on boards who can use [00:16:00] AI from a strategic perspective. Mm-hmm. And, uh, I think that’s easier said. I think that’s easier done I than, than we might think it is. We can get the certification, we could start to use it, just plug and go.

But looking at how can we, uh, transform our value proposition from a strategic governance perspective, I think this big for us in credit unions.

JD Myers: How, if I may, if I may add to that, thank you. AI right now is a great advantage to organizations, credit unions, and it can help move the horizon. Which may be three years out there right now on new strategic objectives, it can help bring that closer, close that window of time on the calendar, closer to six months versus three years, because AI within that organization is going to learn more about that.

Organization’s going to learn more about the context of what that organization does, and it’s gonna [00:17:00] learn about each individual, and it’s gonna learn about the board’s strategy, the board’s vision, the board’s mission, and all the way down through the C-suite. So as you use AI within an organization, it grows with you.

It gains corporate knowledge and awareness, and it becomes one, if you will, with trying to help advance your strategy forward quicker, smarter, with better ideas. So, but that’s a little better. That’s a little better off. That’s gonna take some time. It’s gonna take mm-hmm. Credit unions probably two or three years to reach that point, but it’s gonna happen.

And if a board looks at right now themselves, the board has to take the lead on how to bring AI into the organizational picture. How to design eventually around ai. Mm-hmm. How to develop the organization eventually around ai. So AI eventually becomes the core of many layers of strategy, which eventually ripples out its effects.

Doug English: [00:18:00] Yeah, I really like that jd. It’s like the board was bold enough to reimagine the credit union, like, you know, blank sheet of paper, uh, in the AI area. How would you change your mission? Uh, how, how, you know, would it be expressed differently? You know, people helping people powered by AI or something along that sort of, uh, uh, range.

What do you think, Didi?

Deedee Myers: Well, I’m, I’m thinking a lot of boards go, it’s really hard to imagine the future three years, five years. Why do we do that? That’s why we do business plans every year. What I’m hearing JD say is, if we change that horizon, do some different horizon analysis, AI can bring it forward.

We can still have a vision. This is where we. Need to go to serve our membership. We can’t have that vision, but we can, uh, keep up with the pace of change in a deeper, more relevant way with data, uh, AI powered data, so it can help us look at our products and our services. So I’m hearing a lot about being able to [00:19:00] keep pace, the cadence, uh, really work agility and, um, very robust board conversations, uh, in the future than we have now.

Doug English: Yeah, it, it, it’s a more, it’s a deeper integration is what I’m getting from, from jd. It is like, instead of being, uh, we have an AI project at the credit union. It’s, uh, AI is a part of all of our strategic conversations, sort of another seat at the table of all conversations, looking to build all the proprietary knowledge of that credit union and of its membership and of its strategic, uh, initiatives into our AI model, and then use that to.

To drive the thinking, uh, going forward. And again, like what, what you said in the beginning, ask the AI how to teach the ai, how to prompt, uh, uh, how to, how to use it itself, how to prompt it best, how to go about learning it. Uh, I think that that’s, that’s an interesting vision for the future. [00:20:00] That, that I, I haven’t seen anybody doing it yet, but I guess, I guess you have seen some already.

What, what comments do you have around that, JD or Didi?

JD Myers: Well,

Doug English: the, if,

JD Myers: if I may answer it this way, the board, in my opinion is, is going to need like any other board in any other organization that’s adopting and embracing ai. Need to look at plans as more of a canvas. ’cause AI is gonna provide solutions the board never thought to ask about.

Mm-hmm. It’s gonna be bring very innovative ideas forward on the timeline. So that means your plan, which is in a fixed space, in a fixed verbiage of fixed thought, it may not be able to absorb AI’s suggestions. So we’ve gotta switch to a canvas versus a plan. Canvas is more open, it’s more fluid, it can change the picture with the picture.

And AI is gonna redefine the picture of the CU landscape [00:21:00] almost every month. And so that’s gonna take agility and adaptability with within the framework of board’s strategic decision making. Like you said, uh, the AI component is gonna become a member at the, at the board level as part of the conversation.

I, I see board members during a meeting actually asking ai, what do you think? They’ll, they’ll name it, you, board number 13, whatever. Do they really, what, what do you think of this? And the AI is gonna go well based on the conversation I just heard, but this is what I think. But that on, based on the data, I can cultivate and create knowledge out of within your organization, this is what I suggest.

So it’s, it’s going to be a learning process for the board as well as the AI component that’s sitting at the table. [00:22:00]

Deedee Myers: I’ve got a idea to, to put out here. So when, when boards go look for new board members, often they look for one area of functional expertise. Mm-hmm. In the past it’s been a lot of accounting.

Right. And, and now we’ve, uh, accounting, we’ve got legal, some overrepresented, uh, subject matter expertise. What, what I’m thinking here today is we need board members who can leverage AI and have a broader, more in depth, robust picture of the enterprise as a whole, not just one functional area. How does marketing or branding integrate with our community?

Uh, development effort. What are the products and services doing to impact our balance sheet optimization? I, I think we no longer should be considering just one area of expertise. Yes, we’re gonna come into it with that, but we need to, uh, bring in board members who will welcome the opportunity to learn more about the whole enterprise and, and leverage AI to do that.[00:23:00]

Doug English: I mean, I, I, I, is it a question of, of welcoming them or is it, is it hard to, I gotta think, it must be very difficult, right? These people are super highly compensated if they’re actually in or even near the AI space. Uh, and, uh, getting them to volunteer their time to a credit union, uh, uh, would be a possibility.

But I would think a challenging one. Ha. Have you seen credit unions with success in that initiative?

Deedee Myers: Uh, well, most credit unions will, a lot of the boards will say nobody will join them. Nobody wants to join them. These people don’t have time. Don’t care. I always move with an possibility mindset looking like who does, we will find somebody who has the smarts and the subject matter expertise and the commitment to the community.

We do have to put forth the effort. We need to have a paradigm shift in our mindset on what we need on the board and how we’re gonna find them. Be open to different ways of, of looking and recruiting and get that commitment. But if we start, Doug, like a lot of boards do, oh, [00:24:00] it’s never gonna happen. We won’t get people that’s gonna be out in front of us and a gate and a yield sign, a stop sign.

We constantly have to maneuver around as a filter. I love to go, no, we will find that person and let’s put the effort into it. And with that. It, it will happen. I believe it will happen. And it has, yeah. So we worked with the board the other day and they go, no, let’s go. Let’s shift our mindset. Let’s go find that person.

So we’re actually gonna start a search for them to forget this kind of person on board that will help, uh, increase the level of, I believe, business in acumen across the enterprise for all board

Doug English: members.

Deedee Myers: Uh,

Doug English: I have, uh, uh, AI open on my screen right now, so I could ask it, uh, as the, as the theme is, is what methodologies should you use to, uh, and that, that’s the takeaway, right?

I love the, the vision of, uh, the 13th board member. And that 13th board member is always available. That is the volunteer that [00:25:00] will always work. No cruises needed. They are always gonna be more and more informed. Uh, they’re always going to be, uh, uh, every bit of information that you ever gave it in the past is already there, didn’t forget anything.

It’s got all that there and is ready to help you to try to ask the questions, to take it to the next level. Sometimes it’s truthfulness could be an issue. You need to double check. Uh, yeah. But, uh, what a impact simply adding an AI member of the board. And then the question becomes, if you’re, if you, if it isn’t too technical for the podcast, it, it may be is which AI should be the member of the board.

Any comments around that, JD or Didi?

JD Myers: Sure. At that point, you want an enterprise AI system. So we, as Didi alluded to earlier, you have three, perhaps four types of AI point, AI system, ai, which is a collaborative combination team effort around a certain project, a certain effort, [00:26:00] a certain system. Then you have an enterprise ai, which takes the entire organization into mine, cultivates data, harnesses data from each area within the organization, and comes up with an organizational picture and organizational design for that day on what can be done.

And, and so deedee and I have been discussing this. And we came up with the fourth, uh, idea of where AI fits in. That’s the AI ecosystem. Every core organization has an ecosystem, supply chains delivery, so on and so forth. So AI will eventually reach into every organization’s broader ecosystem vis-a-vis agents, AI agents.

You may have heard about those already. Connector points and data’s gonna flow freely. Knowledge is gonna flow freely, and as it does, each AI system is gonna learn from every other AI system. So the knowledge growth, the wisdom growth is gonna be phenomenal.

Deedee Myers: Hmm. One of the [00:27:00] things that we’re also doing here, uh, is creating a, a position profile, Doug, in an organization, we may call it AI translator or what, but it says if we’re to hire somebody who’s gonna help our organization move forward, help our board govern in this area, time of AI and help management and leadership, you know, what, what would be on that position description.

So, uh, maybe next time we meet, we can share that, but that’s, that’s in process right now.

Doug English: Yeah. If, if I can Go ahead. JD

JD Myers: Antonette very quickly. Thank you. Boards need to understand AI’s already in their organization. So you have the organic from the bottom up of users that enjoy ai. They use it at home, they use it for other purposes, and they bring it into work.

And they’re using AI at work on their own, by their own initiative to, to become more creative with their work effort, wherever that can be done. So you have this organizational growth that is happening [00:28:00] from the bottom. And board members need to be aware of that. They, they don’t wanna damage it, they don’t want to hinder it, they don’t wanna put barriers around it by coming out with rules and regulations and governance.

They wanna, they wanna grow it, they wanna incentivize its growth and everybody else, and it’s going to show. The board members, the executives and the managers. And the supervisors and the frontline team members who really wants to do their work and who really doesn’t want to do their work in the sense of if you put a powerful ai, the board edicts, that’s powerful AI that’s gonna be implemented across the credit union.

And it falls into the hands of people that really are not embracing it. They’re afraid to use it ’cause they don’t wanna lose their job. And I get that question all the time. I bet from employees, is it gonna take my job? That’s their main concern. I said, no, it’s gonna take your tasks, it’s gonna take the tasks that you’re doing that you don’t like to do, nobody wants to do.

It’s gonna take ’em outta your hands, automate ’em for you so you don’t have to worry about ’em. You can move [00:29:00] to the next level of thinking, which I know you’ve got every day you come up with better ideas. And so learn ai, I, I strongly urge ’em, learn AI at home. Bring that value into work. When you do that, you’re bring in, you’re promoting your own value, and, and management will see that the board does need to come up with a strategy for AI, as we discussed earlier.

Doug English: Yeah. I wanna unpack that both of you. I’d like, I’d like you to kind of go, I know you might be creating ideas live here, and that that’s all right. When it’s this front, uh, leading it needs to be created, right? Mm-hmm. Yes. What, any ideas around, uh, or what you’ve seen with the board design for how stray AI is a strategy and how it’s gonna be incorporated?

If you haven’t, uh, heard of some, what might you imagine them to be, or shall I just ask ai?

Deedee Myers: Let’s do both. Ask ai. Absolutely.

JD Myers: Absolutely. The, the, the board needs to be aware [00:30:00] that AI can be dangerous as well as opportunistic, as well as advantage, as well as value. So like any other technology, you’re going to introduce, any other program you’re going to introduce.

Introduce into your organization. There are risks, but risks get overcome, be are overcome by initiators, by innovators, by creators. Creativity people and people that are proactive. People that are proactive will overcome the risks. If it’s measurable and understandable. They can use AI to help them overcome risks quicker.

Or how to close the gap on the risks of AI itself and not, there’s no one single format for, of a framework for any organization. So there’re gonna be multiple frameworks of AI with different learning modules, large learning modules for each big area of an organization, including credit unions. So you might have a separate framework for AI in lending.

You might have [00:31:00] separate framework for AI and investments and so on, so forth. So. These are understandings, components, pieces that boards need to understand. They need to be flexible and aware of that. We may actually be on the precipice of a ca cambian moment in time where it’s such a radical shift is going to take us off in a direction we don’t understand yet.

That’s futuristic, but it could possibly happen. So if boards can come up with, uh, a messaging that indicates very strongly they understand ai, they promote ai, they’re gonna come out with their own AI program. But in the meantime, employees, if you can use AI to, to the advantage of what you’re doing today, please do.

Now. That’s all with within a, a wrap of security. Mm-hmm. We, we can’t let data out that’s sensitive to the credit union, it’s [00:32:00] members to the internet just freely. We have to put security around that.

Deedee Myers: Uh, going simple here. So we’re in October, early October. Almost every organization is in a strategic planning mode.

They bring people together, have conversations. I think it’s a really smart move to make sure AI is in a co in the room. It’s in the conversation. Uh, are we a pilot? Are we a passenger? What’s the role of the board? Are we passengers just coming along with what management brings to us? Do we, do we engage in the conversations and how we wanna use AI to go forward, to better serve the members?

Look at potential merger opportunities, you know, change. Um, our, our growth strategies, uh, I think has to be in the room right now. I think waiting till next year, Doug is not a smart move. And, and it, and, um, yeah, I think, I think we need to be, hopefully they’re bringing this into the room now in some way, having some education.

I [00:33:00] know. Um. I have some clients where we do an education session every quarter, and in November we’re doing a board, uh, governance piece on AI for that board. Like what is the board’s role here? So some of the boards are stepping up and going, what do we need to do here? What do we need to learn? How do we govern?

What questions should we asking? What’s the risk? Um, you know, what’s our, our, uh, governance, governance framework in terms of what decisions to make around AI versus delegating that to the management team. So a whole new conversation.

Doug English: It, uh, it is. And while you were, uh, talking, I asked ai, uh, how can a credit union integrate AI into its strategic plan?

Uh, and after 21 seconds, uh, I, uh, was given, uh, it was number one, start with guardrails. Mm-hmm. Uh, day zero, adopt a risk framework, anchor compliance, early model risk, a third party risk adverse action transparency. [00:34:00] Number two, pick high ROI. Low risk. Use cases, uh, the member experience and the AI co-pilot that JD mentioned in the contact center, uh, operations and risks for document ingestion, uh, lending and collections with controls, uh, and back office policies and procedures.

And we’ve actually had, uh, that subject on this podcast already. Then number three, operating model and governance, uh, governance, the executive sponsor of the AI steering committee. Right. That’s a, that’s a new one that would come out of what you’ve been saying, uh, et cetera, et cetera, et cetera. 21 seconds asking a very basic question.

I think that makes the case for why AI needs to be in your boardroom attending these meetings.

JD Myers: Mm-hmm.

Doug English: Absolutely.

JD Myers: Yeah, I would say so. Uh, if I may add one huge advantage that the board could understand. Make sure they’re aware of, and they [00:35:00] probably are, that the richest data that any organization can have is within the organization.

It’s not out there on the internet. The richest data that any organization can have, they already have it. It’s internal years of it, all kinds of data. Great expanses of data. So what one can do is buy A SLM. It’s a small learning module. Sorry, a small learning module language, and it’s an internal AI system that can be set up with it and it can act like its own AI engine, internal to the organization, internal to the credit union, and have connecting points throughout the credit union.

Have several agents designed to go throughout the credit union and start cultivating data, tagging data, classifying data, attaching labels to data, so it’s not moving [00:36:00] data into one central location, causing a great cost to the organization and great security risk. It’s labeling data, which the AI tracks, and when a question is asked, the first place it can go is within its own organization.

Pull that data together to answer that specific question one might have about the organization and a direction they need to go, or an operational task, trends analysis and so on and so forth in all kinds of ways. And actually ask AI to make a recommendation when it needs to reach out to the internet.

It passes to, it passes that request for more additi additional information to agents that would go out on the internet, grab that data and bring it back. So an SLM is an advantage to an organization, and that’s probably the first. Approach at boards to suggest, should suggest that the entity take is an internal ai.

Doug English: Oh, interesting. Yeah, I have, I have never heard of that before. [00:37:00] Jd. Have either of you seen a credit union doing that yet?

JD Myers: No.

Deedee Myers: Very, very few. Maybe a handful that I know of. And I don’t know ’em all, Doug, but Yeah.

Doug English: Yeah. How would, uh, uh, again, where, where, where is a resource, uh, for our listeners on that kind of, uh, uh, learning more about this, uh, small language model?

Is that right? Yes. Where, yes. Where can our listeners learn more about that ai? Oh, they can go, go ahead. Right. The takeaway from today’s show, just ask ai, right? Yeah.

JD Myers: There’s, this is where, that’s a very good question, if I may, that actually introduces the idea of what’s called the translator. A translator is an individual that understands business speak and it speak and can take the business stack the business operational design, and line it up with the IT stack and the space in between will be [00:38:00] filled with the AI stack and the three shall be aligned.

But this is where a translator needs to step in because it doesn’t understand business. AI business doesn’t understand it’s role in the technology of deploying ai. So a translator would come in and answer that question for you.

Deedee Myers: In, in terms of AI certifications are all over the place. MIT has it, Microsoft, Google, IBM.

We’ve got ’em at, um, a six week one at Stanford. That’s, uh, written about in Forbes, um, Busey, Berkeley. I mean, there’s dozens, dozens ’em out there. I think you have to decide, uh, how long you wanna put into it. What level of detail do you wanna go into it, and you wanna do it from a, a governance framework or, you know, more detailed translator framework.

Uh, there’s some, uh, introductory ones through, um. Uh, IBM, it’s called, uh, AI for everybody that, that’s one, it’s called, uh, [00:39:00] deep Learning ai. So there’s several of ’em out there. I think it’d be advisable for your learning and development officer, HR person to actually put these on your intranet in your organization.

You could also, Doug, have somebody do courses and put them the, uh, master classes on your intranet or your learning management system in the organization. Uh, so I think that’s, uh, something we all need to be looking at, uh, from, from the lower level up and from the board perspective. We could do, uh, look at some resources and put on the board portal.

Uh, that might be more of the six week, uh, shorter timeframe from a strategic perspective.

Doug English: Yeah, maybe a virtual, uh, course, uh, taught by ai, uh, about, uh, the small language model or how to in incorporate AI into your strategy, uh, um, and where to go first, how to build guardrails around it. Like, you gotta wonder, is the AI making sure that these, uh, answers [00:40:00] lead to, lead to its permanence, uh, in taking over society?

JD Myers: Well, that’s a, that’s a good question. It’s, and it’s real, it’s a real question. AI transparency or xai is some calling. It is, uh, a, a part of AI that’s been designed to trans be transparent. So if one needs to step inside of the AI answer and ask ai, how did you arrive at that answer? It is going to lay out the business process of thinking that it went through to arrive at the answer now.

So without

Doug English: now, jd, I gotta, I gotta, I gotta understand that better. ’cause I didn’t think that was an option. I thought the, uh, it was a black box as far as how it got to the answer. You can see the answer, but you can’t see how I, tell me where, how this other version is. Mm-hmm.

JD Myers: Uh, XAI, which is short for explainable ai.

It’s also a vendor term, so you gotta be careful. Okay. Uh, uh, it explains [00:41:00] how the AI started out with the question that somebody asked it, and what process did it go through to answer the question, what data did it retrieve? Where did it retrieve the data? How did it treat the data? Why did it get that data?

And what did it do to converge that data into an answer? That’s a XAI, it’s an part of a framework introduced into. Other frameworks and, and every organization out there is in a rush to, to get to the market, and many of ’em are not including XAI into their model. If you adopt one of those models, you’re at risk, especially if you’re a financial organization, you’ve got to be able to prove to the federal government how you arrived at certain business decisions in certain regulatory spaces.

So if you can’t do that, there’s, there’s a problem.

Doug English: So explainable AI [00:42:00] is something that, uh, like the big company names that we all know in ai, is that one of the options from them or is that a whole different, uh, source of the AI itself? It’s a conceptual framework.

JD Myers: It’s, it’s a thought. It’s, uh, an idea that is being developed out there right now.

It’s not an actual technology.

Doug English: Okay.

JD Myers: Hardware, software, it’s not an application. It’s a construct, if you will, sort of an academic construct that would be embedded inside an organization’s framework in order to peel the onions back.

Doug English: Mm-hmm.

JD Myers: Why did you do this?

Doug English: Here’s why. Here’s the logic behind the research, behind the source.

It’s like the, the source information behind a scholarly article. Like that’s what it’s gonna show you is all those steps. Yeah. That one

JD Myers: of, if I, I’m sorry. [00:43:00] One of the benefits would be. Did the AI start with an authoritative source? Mm-hmm. Or is it going on best Guess data that’s been generated out there by who knows who.

That’s a very important originating step. Where did we get the data from? Mm-hmm.

Deedee Myers: What would help with the governance? Uh, you know, instead of just blindly taking the information that the AI brings to you, you can actually trace what, what were the sources? How did it happen? I think that would help, uh, ground, uh, the decisions that the, the board and the executives are making.

I think also what I’m hearing is it would also translate to increasing or enhancing, evolving our strategic and critical thinking at the board level, as well as the mid-level and lower levels in the organization instead of just blindly being, blindly being a, what I’ll call a passenger.

JD Myers: Mm-hmm.

Deedee Myers: Yeah. So I, I like that.

Doug English: So it’s an, it is a, it is an academic idea, yet it is not yet something you can plug into and use. Correct. So for now, [00:44:00] while we have the, uh, the large language models built in the, the, the predictive, uh, way, you, you still engage with it and use it, uh, in, in perhaps, uh, ways that are not as decisioning, right?

That doesn’t, doesn’t make credit decisions. Is that, is that, would, would that be the sort of, uh, areas to stay away from without being explained? Uh, all the reasoning behind it

JD Myers: from a security standpoint and a viability standpoint? A reputational standpoint? Yes. Correct. Mm-hmm. You know, be careful if you can’t peel that onion back.

Excuse me. If you cannot peel that onion back then, then don’t. Don’t use it for that

Doug English: outcome. If you can’t, if you, if you need to be able to show all the source information around it, don’t use it for that kind of an outcome. Correct.

Deedee Myers: It, the schools are doing that, you know, uh, four or five kids in master’s programs right now, Duggan there, teachers say if you’re gonna use ai, you have to show the source and your thinking and how you looked at the [00:45:00] information so you can use it.

But if you just repeat what AI says without grounding it and backing it up and validating it, you’re outta luck on that assignment grade. So I’m, that’s, we’re learning it in university, so that’ll be helpful.

Doug English: Yeah, it’s, it is, uh, I was, uh, teaching a couple of our, uh, staff, uh, some, some people that were trained to be certified financial planners.

So I was doing some time value of money calculations with them. And I started out by, uh, asking our friend, uh, AI to come up with the questions. I did a brilliant job of that, and then I asked, uh, to give me an answer sheet, and then I checked the answers and it was completely wrong. Absolutely beautifully convincingly.

Completely wrong. Uh, and it just shows you that, uh, that that source information and checking to see accuracy is still very real. Even if something is basic as just a formula, it, it cannot be relied upon for, uh, for complete accuracy. But its ability to idea generate is truly [00:46:00] extraordinary. And the ideas that you can come up with in the boardroom, in the, uh, in the strategy room, uh, for your credit union can be just massively enhanced by putting that 13th member in welcoming AI into your boardroom.

Uh, and then, uh, building a framework around how to let it be a part of your strategic process in a well defined, carefully risk controlled. Matt is my takeaway from talking to you. Very smart people today. So, with that said, any other takeaways, resources, ideas that you wanna offer to our listeners for how to take this, uh, this idea to the next level beyond the obvious one?

Just ask ai. Just ask

Deedee Myers: ai. I would encourage, uh, executive teams, mid-level supervisors, directors to maybe have lunch and learns and somebody puts a topic up and everybody learns together using AI and from different perspectives and see what comes out of it as a practice. [00:47:00] Uh, that’s very easy. It’s cost.

Cost effective really doesn’t cost anything except time and a pizza to order for lunch. But start, start there by modeling that we do embrace, uh, the thought of AI or learning about AI and embedded into our learning and development just through a lunch learn. It’s. Sounds easy, Doug. I think we can do that at the board level as well.

I’m also have, through my work, um, looking at AI from a board governance perspective. Got a little scared. I saw some stuff out there about when I researched AI on what’s the board’s role in strategic planning in a credit union. What came back to me, I totally disagree with, but I’ve heard other speakers speak and I think we’re, we can misdirect our board members, uh, by using AI without going back to the source of what true governance is.

Good governance is. So, uh, caution flag there, uh, went up for me.

Doug English: Hmm. So be [00:48:00] careful. Don’t take, don’t just take sort of my example as well. Don’t just take it as fact. Check it. Think about it. Does it resonate with your strategy? Who you are as an organization, where you’re going? And good governance,

Deedee Myers: not just ’cause somebody said it, it’s out on the internet now, it’s an ai.

Um, yeah. So we’re, I think we’re, we’re being a little careful there,

JD Myers: JD another. Sure. Um, ideally the best way to start off with AI is to take the vendors, have a discussion with your current vendors, whatever vendors you have in your organization, whatever software vendors you have, big enterprise systems, small applications, and ask them, are they integrating AI into their updates, their products?

And if they are. They’re, they should notify you. They should notify the board members, the operators of the credit union. What is it they’re in, integrating in the way of ai the capabilities [00:49:00] and programs and so on and so forth within their existing software, enterprise software, uh, applications that they have within the credit union.

Most of the time they’re not the best. They, uh, are actually quite unimpressive, but over time, the vendors will learn how to better integrate. What type of AI to integrate, how to tune that AI to the organization? There is a learning process for the vendors, and the vendors are in a learning process just as the credit unions.

Mm-hmm. So there has to be flexibility and understanding, cooperation and bring down the expectations a little bit, make them more realistic and working together to, uh, start integrating AI in, into the credit union. I would suggest that working with the vendors and their embedded AI is what it’s called, versus trying to bring in and introduce a whole external AI system that could actually conflict [00:50:00] with the existing systems that are in the credit union.

So it has to be an integrated approach. Well planned out and proof, some proof of concepts along the way.

Doug English: So jd, are you suggesting that the credit union look at the systems that the vendors are using because the credit union itself might use that same system? Is that the, the logic,

JD Myers: it’s, it’s, for instance, like Salesforce, if the credit union is using Salesforce, to what extent is the current day Salesforce in that, in that organization?

To what extent is it using AI within Salesforce? Is there an embedded AI component within Salesforce? That’s the question that needs to be asked first. Uh, any CRM system, any ERP system, enterprise resource planner, customer resource management system, likely has embedded AI in it today. And so the question would be how is that going?

[00:51:00] What’s the, what’s the. Trend line for that AI within those systems and what it can do, what it, what it can, its cap capacity and capabilities, what can it deliver and have that used by the organization.

Doug English: And then is the, is is the next step where the credit union considers that same AI for other uses or because I know you, you were saying as compared to if you just go out on your own and select something, uh, looking at what the vendors are, are using is a, is maybe a better way to go about it.

I’m just trying to figure out what’s the next step beyond that.

JD Myers: So let’s say you have three major vendors within the credit union. You have Salesforce, a separate one is a customer resource management system, or customer service management system, and the otherwi other one is an enterprise resource plan planning system or program that takes care of the enterprise a as a whole.

Does those three, do those three systems have [00:52:00] embedded ai. Working that would serve as an advantage to that particular credit union that’s using them. Mm-hmm. And so the question will likely be, yes, we have embedded ai. Yeah. Okay. To what extent is that embedded AI being used to provide the answers that we’re trying to get out the system every day?

Right. Okay. So the customer needs to be updated and made aware of constantly, not constantly, but through iterations of the power of the embedded AI that’s in the system, because that might help the credit union board make a decision on what type of external AI they need to bring in. Okay. That’s not part of those systems.

Right.

Doug English: Got it. Yeah. And, and now I understand now that, that it is, it’s, it’s shortening the, uh, the decision tree, the number of things that credit union might consider for their own initiatives. Got it. Very good. Awesome. Well, uh, thank you for being, uh, so bold to join me for [00:53:00] this, you know. Topic that we haven’t all figured out yet, right?

Like we are trying to figure it out. And that’s the purpose of the podcast, is to come up with the bold ideas to help power the credit union forward, the credit union movement forward. So thank you so much, DEI, for coming back again. And jd, I, I love the, the topic that we’ve discussed. I don’t think we’re done with it.

I think we might need to do this right a couple more times. Uh, uh, and I will, uh, look for your leadership, uh, and thought ideas, uh, as I always have before. Thanks so much, everyone. Thoroughly enjoyed it, Doug. Thank you. Take care, jd. Bye Dede. Bye

bye.