Is the AI Adoption Gap Caused by a Curiosity Gap?

Recently, when scoping out a project with one of our clients, they gave us an unexpected definition of success: curiosity.

They have a team of relationship managers (RMs) who run high-touch annual workshops — arguably the most important client touchpoint of the year. The workshops need to feel personalized and dynamic. But the RMs are spread across dozens of accounts, and going deep on every customer takes the kind of preparation that only happens when someone is curious enough to dig.

"If we make the workshop feel much more personalized and dynamic, the customer will enjoy it more."

Then it got interesting:

"To make it more dynamic, our RMs need to be more curious about the customer — the company itself, their industry, the team members attending the workshop, etc. — because if they're more curious, they will go into more depth when preparing and running the workshop, which will make it richer with context and more realistic. And all of that will create a better experience for the customer."

So much of AI chatter is about automation and efficiency, but our client saw something else: using AI as a way to change a personality trait. But could you really make someone more curious?

We reframed the question. We couldn't make someone more curious. But we realized we didn't have to. AI could do the work that curiosity would have driven — the deep research, the digging into each customer's context, the preparation that a naturally curious person would have done on their own — and hand the RM the output. The RM doesn't need to become more curious. They just need the benefit of what curiosity would have produced.

It worked. The workshops got better. But building it raised a question we couldn't shake.

If AI can produce the output of curiosity, can it also produce the impulse?

The pattern we keep seeing

We keep having the same conversation with friends and colleagues — people across every kind of role. When AI comes up, they say some version of the same thing: "I know AI is so powerful, I'm just not sure how to use it in my job."

We always ask the same follow-up: "So why don't you just ask AI how to use it in your job?"

Same response, every time: "Well... I never thought about that."

These aren't people who lack capability. They have the same tools, the same access, the same ability to type a question into a chat window. Tech-wise, capability-wise, there's not a real difference between heavy AI users and non-users. The gap seems to be entirely psychological.

Barriers to curiosity

Wharton's Human-AI Research group published a blueprint on AI agent adoption after interviewing executives at companies deploying AI at scale. They identified three psychological frictions that determine whether someone will actually use AI:

1. Perceived competence — Do I believe it can do my job?
2. Trust — Can I trust it to act on my behalf?
3. Delegation of control — Am I comfortable handing over any part of my work?

These are real barriers. And in the three-plus years since ChatGPT launched, many people have hit one or more of them. They've tested AI, found it lacking, and walked away. They've felt the competence doubt, the trust gap, the discomfort with handing over control.

But there's another group just as large: people who never started at all. Gallup's data from April 2026 found that among workers who have AI tools available but don't use them, 46% say "I prefer to keep doing my work the way I do it now." Roughly 40% don't believe AI can be helpful for the work they do. Not afraid. Not distrustful. Just absent.

What do both groups have in common? Neither one is curious enough to push through.

We think there's something upstream of all of this that the research is missing: curiosity.

The people who tried AI and bounced off it? Curiosity is what would have pushed them past the wall — to keep testing, keep experimenting, keep asking "what if I tried it this way?" instead of stopping. A curious person says "Maybe AI isn't competent enough, but let me see..." A less curious person tends not to.

The people who never started? Curiosity is what would have gotten them to open the chat window in the first place.

The absence of curiosity lets fear and indifference win by default.

This reframes the problem. The frictions Wharton identified are real, but they describe what happens after someone has already engaged with AI. Most of the energy in AI adoption today goes toward helping those people — building trust, proving competence, easing the handoff of control. Meanwhile, the people who haven't started at all get a mandate and a login. Companies are building training programs for people who aren't curious enough to attend. They're creating change management strategies for people who don't think any of this applies to them. They're solving for frictions 1, 2, and 3 while ignoring friction zero.

Fear without curiosity is paralysis. Fear with curiosity is progress.

Can one prompt do the trick?

With our client, AI substituted for curiosity — it did the deep work so the RMs didn't have to be curious themselves. But that required us to build the solution and put it in their hands. It doesn't help the person sitting in front of a chat window who doesn't know where to start.

For everyone else — the afraid, the indifferent, the ones who haven't started — you need to spark curiosity itself. That's the gap. And we think a prompt might be the bridge.

Leaders are investing in AI by buying tools, adding licenses, and rolling out AI features in existing software. But most can't show their people what to actually do with it. Gartner found only 26% of executives are confident and proficient in AI themselves. The mandate is clear; the guidance isn't.

But if everyone already has AI chat, then maybe they just need a structured first step — the exact words to start a conversation with AI about their own work.

We iterated on a prompt over dozens of versions — using AI the entire way, from brainstorming to planning to writing to testing — and built something designed to do exactly this. Here's what happens:

Someone pastes the prompt and hits send. The AI greets them, explains its role, and asks them about their actual work — what they do every day, what takes up their time, what frustrates them, what they're good at. They respond with as much or as little detail as they want. Then the AI offers a handful of specific, tailored ways it could help — not writing their emails, not generating slop, but ideas that augment their strengths, bring a novel approach to their work, or take the tedious parts off their plate. Whatever they pick, the AI works with them on it as a collaborator, building trust while showing off its skills and keeping them in the driver's seat.

That's the moment curiosity activates: when someone sees something specific to their work that they hadn't considered. They go from "I don't know how to use this" to "wait, that's interesting." And once that door opens, it tends to stay open.

You can't mandate curiosity. But you can mandate a single action that triggers it.

Curious where this goes

AI is widely accessible, but adoption at the individual level is still lagging. We think it's a curiosity gap, and that a personal, low-friction AI conversation is the bridge that can get people over it.

Now we just have to see how it works in the wild. Hopefully there are some people curious enough to try it.