AI Only Tells The Story You Give It

Joe Winn, President, GreenProfit Solutions

Three credit unions, no birthday messages. Yet those same institutions aren’t shy about sending “boy have we got an insurance offer for you!” notices. By email and paper envelopes. They know who I am. They reach out when it benefits them.

We all read way too much about how AI is changing the game in loan processing, fraud detection, marketing content, and member service, to name a few. As a member, I’d prefer it to be invisible, leading to systems with less unnecessary friction or frustration and more moments of delight.

But those last two are, by nature, visible. Generative AI tools are arriving in those environments whether you’ve approved them or not. The natural instinct, to use said tools for efficiency and speed, makes sense, but we’re already ahead of ourselves. There’s a problem we skipped:

What does “more, faster” actually produce? Is the underlying identity aligned with your mission and consistent across all channels, services, and audiences?

AI is an amplifier; if your “instruments” aren’t all in tune, you’re not going to get an epic symphony when you turn it on.

Bohemian Rhapsody didn’t happen because Freddie Mercury turned on an amp and prompted, “genre-bending song; make it ridiculous and historic”. Instead, AI tools produce more of what you give them, with tweaks detracting as often as they help.

If your brand voice is clear and your community story is consistent, these tools help you express it more efficiently. However, if that voice is one thing in phone trees, another in email blasts, and totally different in person, then…

Congrats, you just bought cheaper fragmentation.

Say you wanted to ensure a new platform retains your credit union voice. So you feed it the past few quarters of emails, newsletters, and website copy. Seems reasonable, right? It’s a great first instinct. But then:

All the results, no matter your prompts, sound like they were rushed through a Friday evening committee meeting. It’s…fine, but best case, your marketing people will have to edit to be up to your standards. At worst, you just remove some em-dashes (we’re being honest here) and make it live.

The tool did what you asked: It took an inconsistent voice and mashed it together into a cacophony of words and formatting, literally desperate (I’m personifying a non-sentient tech, but “make you happy” is high in its system priorities) to please. A multi-billion dollar mirror.

Let’s ensure the reflection reflects the passion, care, and meaning your institution exists to provide, specific to the goals and challenges you want to address.

Big banks have an assumed capability to provide any needed service. After all, isn’t that why they’re big? (Not necessarily, but also not the point of this article.) Fintechs have character; they were built identity-first, with the product following the brand design.

Meanwhile, credit unions arrived mission-first, assuming their identity out of it. This made sense with the 14 people who initially came together, and was still relevant when they represented the residents of a banking desert, or the members of a labor union. You were your identity.

That’s often not the case anymore. “People helping people” is a value, not a voice, especially when it’s shared across 4,400 credit unions nationwide. It’s still important, but it isn’t who you are, it’s what you believe, an inherited (and rightly shared) concept.

For a credit union to make AI tools speak/look like your institution (across chat, email, marketing, everywhere you plan to use it), and actually be a net positive for effort, trust, and connection, you have to clearly define your voice.

Let’s check your own. Pull the last five pieces of member communication your institution sent out. Would a 32-year-old who just moved into your community understand what the institution is at its core, why it exists, and whether any of that matters to their financial life?

Can they describe what sets your credit union apart from Chime, Chase, or whatever institution they use now? And is there an emotional hook which might drive them to at least learn more?

If your answers to these questions are mostly or all “yes”, then you’re good…genAI can likely save you time and effort. But if, instead, it comes across as multiple institutions all showing how they can help in a different key (I’m stretching the music metaphor hard), then that’s the thing to address first.

Not just for training generative AI tools, but also because existing members who received those communications are already experiencing that disconnect.

The credit unions that rush to implement “AI everywhere” because it’s “the future” will only automate their inconsistent image. Those which thoughtfully incorporate AI tools will reap the greatest benefits, enabling it to enhance, not replace (or worse, confuse) their message.

There’s a clarity problem with some credit unions, and while it is expressed in the technology and seen in the marketing, it’s actually an identity issue. Getting on top of it before AI further muddies the water of what you stand for is essential.

The good news is that it’s solvable, but the window for solving it is closing faster than many leaders realize.