“Is it really human today?”
That question about member services hung in the air during a conversation between Sarah Snell Cooke of The Credit Union Connection and Glia CEO Dan Michaeli at GAC 2026. Credit unions often worry that artificial intelligence could strip away the personal touch they pride themselves on. But Dan flipped that idea on its head. When you compare the promise of AI to the reality of today’s member experience, he suggests the bigger question might be whether the system credit unions rely on now is truly personal at all.
At the center of the discussion is a simple ideal: AI as an extension of the credit union team. Not a replacement for staff, but a digital workforce that supports members, coaches employees in real time and helps leadership understand what’s actually happening inside thousands of daily conversations with members. It is a vision where technology quietly clears away the noise so people can focus on relationships.
One of the most intriguing moments in the discussion comes when Dan describes the untapped goldmine credit unions already possess. Every member call, chat, or video interaction contains insight about what members care about, what they struggle with and how they respond to products. Historically, most of that information simply disappears after the conversation ends. With AI, those everyday interactions suddenly become a living source of intelligence.
The conversation also touches on something credit unions wrestle with constantly: capacity. Frontline teams are often overwhelmed by routine requests, leaving little time for proactive outreach or deeper relationship building. Dan suggests that when AI absorbs a meaningful portion of those repetitive interactions, staff are no longer stuck reacting all day. They gain the breathing room to do the very things that make credit unions different.
And perhaps that’s the most interesting thread running through the interview. For all the fear surrounding AI, the technology might actually create more opportunities for the human moments credit unions value most. Not fewer.
NOTE: The AI responsible for this masterpiece should probably stay in its lane—which, based on this output, is not transcription.
Sarah Cooke
I’m Sarah Snell Cooke, host of The Credit Union Connection. We’re live here at GAC 2026, and I’m joined by Dan Michaeli, welcome. Thank you so much for having me, and he is the CEO at Glia, tell us a little bit more about yourself in the company.
Dan Michaeli
Absolutely. Well, Glia is a banking AI workforce, so think about it as the number one platform for intelligent banking interactions. The idea is, we want to provide an extension of your team, right? So think you know banking AI that works across members, your MSRs, your managers and your executives in different capacities. So that could be a banking AI agent, right that works directly with members. It can be coaches and copilots that work with the MSRs to make them more effective, more efficient, provide a better experience, and then AI analysts that work with your managers and your executives across all the different channels, so across voice and digital, any channel, regardless, it’s an extension of your team.
Sarah Cooke
And so everybody’s talking AI. What I want to start a little bit broad. Where do you see AI going and where do you see cranes going with it and say, five years from now?
Dan Michaeli
Well, I believe that what will we what we will be seeing is exactly sort of that extension of the team concept, right? So ideally, what you will see is the ability to use AI in order to further the Credit Union Mission in order to connect much more deeply with members, because today there’s so much volume of interactions that comes into the typical credit union that takes up so much of the time, and ultimately is not yielding the experience that you Want, right? And so the idea is that you can leverage these capabilities to deepen relationships and actually further the differentiation. So differentiation for credit union is that member connection, right? So in order to do that, you have to achieve a different cost structure. It has to, you have to get to a different level of capacity within the institution so that you can deliver that and that that capacity is what AI delivers.
Sarah Cooke
And so you’re talking about an MSR who can find an answer faster because it’s right there, and then get the member that what they want faster. And so they’re not being impatient with you, and the employee themselves has a better experience too, because they’re not frustrating, not able to find things.
Dan Michaeli
Absolutely I’m talking about yes. Not only, not only can they find an answer that they may be looking for faster, but they can also receive suggestions in line in real time. If they’re on the phone or in chat or on a video chat, they can get coaching suggestions right there in line with the conversation, right? Similarly, if you’re a credit union, an executive or a manager of the MSR team, right, you can, in natural language, ask questions about what is happening across all of these different channels and all of these different conversations, and get answers and example interactions. How is that CD product being received by the base right? And you can see immediately, just with that natural question, you can see examples, charts, tables, graphs. You can see sample interactions of what members are saying about it. So just a whole other level of depth, because we can tap into all the conversations that are happening, not just the knowledge base, not just the documents, not, not just the sort of internal documentation, but all the conversations that are happening every day with members.
Sarah Cooke
And that’s one of the things I think is really interesting. Because 10 years ago, it was, get your data clean, get your data together, blah, blah, blah, because it was in 50 different places, and sometimes it’s still is but, but then actually taking that data and using it for something. Because sometimes people have the data, you know, what the heck to do with it? And it sounds like it’s a it’s your business analyst, which and so you’re not paying $300,000 for somebody to do that.
Dan Michaeli
Well, exactly right? And you know the act, it’s so interesting, if you think about the amount of data that exists in the conversations that are happening every day. I mean, yes, it’s valuable to have all of these documents across the organization that outline policies or procedures or compliance requirements, but these are this information is in the conversations that are already happening. So if you have all of that unstructured data already happening across the conversations through all these different channels. What if you could access that, right, and not have to worry so much about cleaning up your data because it’s in real time, it’s happening every single day in all these conversations, right? So that’s what’s so magical about the unstructured aspect of AI, where you can tap into these data sources without having to worry. So much about just cleaning it and keeping it so precise. You know, it’s much more accessible.
Sarah Cooke
And I mean, to my mind, the biggest threat against credit unions is the lack of brand awareness of what they are. I mean, some credit unions, individually, yes, do very well, but as a whole, on the industry, you go ask people out on the street, they’re not going to know what a credit union is nine out of 10 times, I’m sure. And I think something like this is probably going to help credit unions better tell their story, to build that brand. And I’m just curious, what you know, do you have credit unions using at that capacity as well? In addition to, like, not the business analysis, but like, Hey, give me all the positive feedback we got from members on the phone today, whatever.
Dan Michaeli
Yes, that is a that is a great use case. Look, ultimately, our purpose as a business is that we want to help build thriving communities through banking interactions. That’s how we think about what we deliver to the movement, right and empowering them to fundamentally change the amount of staff that they need to handle interactions, because then all those people that are have been, let’s say, in the just in the frontline team, right, but across all the different teams, making them more Effective at their work, right? Allows you to take additional staff and reallocate them to other initiatives. And that could be going outbound to, you know, potential memberships, or that could be calling indirect, you know, folks that that have indirect loans, calling those folks and trans, transforming them to full membership, right? Right? So it the ability to have be more front footed, right is, is right now limited, because there’s so much everybody feels that they’re just underwater. You know, the frontline teams just feel they’re underwater. They don’t have enough staff, and then people leave, and they have to train new people, and it’s just a never ending treadmill, right? When you eliminate that using AI, all of a sudden, the people that you have on your team can focus on what makes credit unions so great, and I’m promoting that brand, and I’m doing the outreach and sharing, what are the compelling benefits of being a part of a credit union, right? What are the what are the benefits of becoming a member? Right? Right? Because that’s where the differentiation is. That’s where the magic is. What is your differentiation at your credit union? Exactly, and being able to tell that story, having the capacity, your rates, it’s the it’s the connection that you have to the community that you serve.
Sarah Cooke
Yes, and I think, you know, during covid, it was kind of called to the credit unions are kind of called to the carpet, because they called to the carpet because they didn’t have their tech in place to serve and that’s when they started lowering the member satisfaction below where the big banks were even and it sounds like this is something that is going to spin that back up again.
Dan Michaeli
It really is. You know, it’s ironic, because a lot get a lot of a lot of credit unions are worried about AI sort of having an impact on that personal experience that they can provide, right? And then what’s so funny is that I will, I will often sit across from a credit union and will call their member servicing line, right? And we’ll hear what the experience is today, right? It’s an antiquated telephone banking experience. It’s never, I get a human right away, right? It’s always had, there’s always a press one for this, press two for that. There’s always like, dial an extension, or if you want your balance go through an authentication dot, you know, use this. Use this. You know, old school telephone banking system to get a balance check, right? These systems are in place today, right? And so ironically, when I compare that to, well, what if we could have a conversational experience where I pick up and say, I say, Hey, David, you’re calling back. Thank you for calling back. I know that last time you were calling about your direct deposit or a fraud inquiry that you had or, you know, you were actually calling to check your balance. Is this any of the topics that you’re looking to discuss today, right? And that’s a totally different experience. And once that interaction actually gets to an MSR, that MSR isn’t underwater, because the AI is helping them handle all of these interactions at the front so they have so much more time. Some members just want to call to have a conversation, right? Wouldn’t it be wouldn’t it be wonderful if we could we didn’t have to rush them off the phone, right? Wouldn’t that create the better member experience? So when you actually look at them and put them side by side, I believe AI is the unlock to creating the most personal credit union, credit union experience you can possibly have.
Sarah Cooke
Yeah, humanize it with technology, absolutely. And I understand glia did some research consumer behavior and credit union service and things like that. So tell. Us a little bit about what you found there.
Dan Michaeli
Yeah, you know, it’s, I recommend everybody going to our website and downloading the report. It’s our benchmark, annual benchmark report where we look across all of our clients that are leveraging these capabilities and bring you the most important insights that we that we are seeing, right? So these are real deployments in production that that are teaching us so much about the way to bring AI to the credit union movement. We have over 400 data, 400 different credit unions that participated in this report. And you can see, for example, specifically, which are the interactions that are intentionally being sent to an MSR, right, or which are being escalated as part of the interaction to an MSR. So intentionally means, hey, look, I this. These are the different types of these are the different types of topics that we’re looking to cover. And I every time there is a question about, you know, a specific product I want that immediately sent to a person. I don’t even want the AI to deal with that. So there’s an intentional approach of sending a conversation to a an MSR. And there’s also the escalation, where maybe the AI doesn’t get to a point of understanding what the what the member needs, and then it’s escalated to a person, right? The world class level of understanding that we see with our AI is over 92% understanding. So the REI is understanding members questions, because you can say, you can ask a question in infinite number of different ways, right? So the understanding rate is so essential, right? Because you have to identify what it is, what is it that they really are looking to solve? We understand 92% right? And then the decision is, do we want to once we understand it? Do we want to keep that, keep it with the AI, or do we want it to be sent to a human right? So in the report, you can see a table that summarizes the top types of topics that are that are handled by AI, how are they handled in best practices by all of those 400 credit unions? It’s really very insightful to see actual data, because there’s so many failed AI projects out and stuff that is, it isn’t in production, stuff that’s vaporware, stuff that’s not real. You see a demo, it’s like that looks really fancy. But do you have any production clients? You know, we have three times the number of production clients than our nearest alternative solution, right? So we have an incredible amount of data, and that’s all packaged up in this report. So I don’t want to give too much away so that people go and download it.
Sarah Cooke
Absolutely. It’s some really insightful stuff, yeah, and so we’re in Washington and for the Government Affairs conference, and there’s going to be regulation and legislation around AI. More, yeah, more. So what are you anticipating might come in that realm, and how are you all planning to handle it so the Ukrainians can continue better serving their members using the AI.
Dan Michaeli
I think one of the most important things that we need to look at when it comes to regulating AI, or understanding how to deploy these projects safely, is when to use generative AI and when not to use generative AI. The ultimate combination is, or the ultimate approach is a combination of generative and deterministic. AI, right? Because we are a regulated industry, we cannot just open up and say, Oh, I’m going to put some guardrails on this and it’ll all be okay, right? The problem with this approach is that there are an infinite number of ways that these systems can be hacked. So it’s not a question of if, it’s a question of when, right? So anything that we look at when it comes to these capabilities, we need to make sure that we’re not opening up that risk, right? And so that’s where I think that we need that’s where the regulation could come into play. Is just when and when not to accept uses of generative AI. Our system combines both. It brings the best of both worlds together so that our clients get the cutting edge aspects of generative AI, but the security of deterministic AI, right? And so I think that the balance between those two things is an area where I think it’ll be, it’ll be interesting to see how the regulation pans out on cyber security.
Sarah Cooke
Isn’t bad enough? But, yeah, no, I think you’re, you’re dead on because we there’s gonna need to be guard rails. I’m sure the government wants to put its fingers in there. I so I always allow my guests to have the final thoughts. What would you like to leave our credit union audience with today?
Dan Michaeli
Well, I would, I would encourage you, your audience, to check out the report, because there’s a there’s so much information out there that isn’t actually rooted in. Live production deployments at scale that are producing ROI AI with art without ROI is a toy, right? It’s not a real it’s not a real business transformation initiative if it doesn’t have ROI attached to it, all of the clients that are contained in that report have incredible stories to tell. And for those of your listeners that are thinking about this, which I know all of them are at this point, right, they, they, they should. They have access to real, live data and takeaways from the early adopters to learn how they can bring these capabilities into their organization and the last thing I’ll say is just think about, when you think about how much more personal you can make your experience with AI. Don’t really put the experiences side by side and really think, what is the member going to experience today versus if I deploy AI? Because there’s this again. There’s this tendency to say, well, I don’t want to remove the human touch. But is it really human today? If we look at it, is it really what you want to create today? And if the answer is, it can be better, that’s where you really have to start thinking about how to bring these capabilities in excellent Well, thank you so much for your time. Appreciate it pleasure. Thank you for having me.