Your AI Strategy Is Really Your Data Strategy

Lately I’ve been writing about the forces pressing in on community banks, from deposits quietly moving onto crypto rails to fraud that keeps getting harder to spot. AI sits underneath all of it, and it may also be the part of the conversation you’re most tired of having.

I get it. There are only so many vendor demos and think pieces you can sit through before the word itself starts to feel like wallpaper.

But the reason we can’t stop talking about it is the point. When an entire industry can’t look away from something, that’s usually worth paying attention to. The American Bankers Association surveyed more than 250 bankers on exactly this, and the picture it paints is one I recognize well: an industry that knows AI matters, moving carefully, and not always sure where to start.

That last part is what I want to talk about.

In my experience, the question community banks are asking, “Where do we begin with AI?” is almost always pointed at the wrong thing. They’re looking for the right hire, the right consultant, the right governance framework. Those things matter eventually. But they’re not the starting line.

The starting line is your data.

What the Research Is Telling Us

When the ABA asked bankers what’s blocking AI adoption, the answers were consistent: lack of expertise, lack of a clear business case, lack of internal buy-in. I hear versions of this constantly, and I don’t doubt any of it. But I’d push back gently on the diagnosis, because those aren’t the root problem so much as what the root problem looks like once it reaches the surface.

Only 3 percent of banks in the ABA survey said their data infrastructure is optimal for AI, and nearly half described it as only partially structured. When Ron Shevlin measured how well banks actually use the data they already have in the Data Execution Quality research we commissioned, the average score across 100+ institutions was 241 out of a possible 500, a little more than halfway to where they need to be.

You can hire expertise and still get nowhere if the data underneath is unreliable. You can build a convincing business case on paper and still fail to execute it if your information isn’t clean, accessible, or trusted by the people who are supposed to act on it. The “lack of” findings in that survey are real, but they’re symptoms, not causes.

Ron put the cause about as plainly as it can be put: there is no AI strategy without a Data EQ strategy. The research goes a step further. When a function’s Data EQ score falls below 60, the AI efforts tied to that function don’t just struggle, they fail, because the technology can only ever be as good as the data feeding it.

Here’s why that matters so much. Point AI at broken or fragmented data and, rather than closing the gap, it widens it, acting at speed on whatever it’s given. We’ve watched a version of this play out in analytics for years, and the stakes with AI are considerably higher.

So Where Do You Start?

I’m not going to tell you to hire a chief AI officer or bring in a big consulting firm. What I’ll tell you is that the institutions making real progress right now are the ones that got honest, first, about what their data can and can’t tell them. Here’s where I’d focus.

• Build one source of truth your leadership team trusts. If your executives are working from different numbers, or waiting days for a report to land in their inbox, the foundation isn’t there yet. This isn’t an IT project so much as a leadership decision about what your bank measures, how, and who can see it. Everything else sits on top of it.
• Take stock of what you’re capturing, especially the parts that never make it into a spreadsheet. Transaction data matters, but the context lives in the qualitative layer: loan officer notes, call center conversations, customer feedback. That’s the raw material large language models will eventually do their most useful work with, and the Data EQ report calls it a goldmine most institutions don’t know what to do with. Most community banks aren’t capturing it consistently today, and building that discipline now, before you need it, is one of the quietest, highest-return moves you can make.
• Have the culture conversation, not just the quality conversation. In that same research, 72 percent of high performers said they treat information as a strategic asset. Among low performers, that number was 3 percent. The gap between those two groups is less about technology than about whether leadership treats data as something shared and strategic, or something that belongs to a department and gets pulled on request. That tone has to come from the top.

The Institutions I’m Most Confident About

They’re not always the ones with the flashiest AI pilots. They’re the ones where a CEO can tell you, without pulling up a report, how their deposits are moving, where their loan officers are performing, which customers are showing early signs of risk. The data is alive in how they run the bank, every day. For those institutions, AI is the next chapter of something already in motion. For everyone else, that’s the chapter worth starting now.

So here’s where I’d push you, and I’d push hard. The AI clock is already running. Cornerstone’s research shows roughly a third of community institutions have already deployed chatbots and a quarter are using generative AI, and the gap between the high performers and everyone else widens every quarter. Don’t wait for the perfect use case, the next budget cycle, or the vendor who promises to make it all simple. Get an honest read on your Data EQ, function by function, and fix what’s holding you back before you point AI at it. The institutions that move now will compound that lead, while the ones that wait spend the next two years explaining why their pilots underdelivered.

I’d be remiss not to add that we feel this at KlariVis too. We’re putting real work into the AI side of our own platform, held to the same standard I’m laying out here, because we never wanted to have this conversation from the sidelines. We understand the pressure you’re under, and we’re building to be better for the banks and bankers we serve, with more to share soon.

The full Data EQ report, which Cornerstone built with us, lays out exactly how to benchmark where you stand, and you can read it here. It pairs with what we’ve been showing in our deposit and fraud research: the banks that can see their own data clearly, in real time, are the ones ready to act, whether the pressure is stablecoin outflows, fraud, or an AI race they can’t afford to sit out. The data is already in your bank, and the only question left is whether you move on it before your competitors do.