Why Two People Use AI All Day and the Other Six Don't

On many small teams, two people use AI all day and six don’t. You can probably picture the two on your team without thinking. They’re the ones who reach for ChatGPT, Gemini or Claude for almost everything. The other six are harder to see clearly, which is the part most AI champions miss. The instinct is to read this as interest, or maybe natural aptitude. Across the small teams we’ve watched, that almost never turns out to be what’s going on.

What the split actually is

The two-and-six shape shows up almost everywhere. Marketing teams, operations teams, professional services firms, finance departments, in-house support groups at companies of all kinds. The functions are different, the seniority spreads are different, the tools the company has paid for are different. The ratio holds. That’s the tell. A skill gap would not land the same way across that much variation, and neither would interest.

Something structural is doing the work.

Most of the energy aimed at closing the gap aims at the wrong layer. Companies buy everyone a subscription, send the team to training, run a lunch-and-learn, then watch the original two get faster while the other six stay where they were. The premise is wrong. You can’t push more inputs into a system whose real bottleneck is somewhere else and expect the bottleneck to move.

The two who use AI all day share three quiet conditions the other six don’t, and none of the three are about the tools. Leaders looking at the same picture from above tend to read it as one of several adoption gaps to fix.

From inside the team, the split is the whole story.

What the two share that the six don’t

The first condition is that they’ve watched their manager actually use it. Not endorse it. Use it. There’s a real difference between a leader who tells the team AI matters and a leader who actually uses it in front of the team, and the two are often not the same person. The six are reading behavior, not policy. Gallup’s 2025 workplace research finds the same pattern at a much broader scale. Where managers are seen using AI rather than just talking about it, their teams follow. Where they only talk about it, the team doesn’t.

The second is that they have somewhere to start. A blank chat box is intimidating when you haven’t done this before. You have to invent the question, write it well, and hope the answer is worth the time it took. The two don’t start from empty chat boxes. They have an example from a colleague, or a version that worked once and got saved, or a short routine they figured out and reuse. The six are still typing into the empty box every time, which is why each attempt tends to feel like more work than just doing the thing the old way.

The third is that they’ve watched someone with their job get a real result on their own work. Not a polished demo. Not a webinar. The kind of task they were going to do anyway, run by a colleague, producing an output they recognized as actually useful. The “person like me” piece is the most underweighted of the three. The reason some people decide AI isn’t for them usually isn’t the tool. It’s that they’ve never seen the tool used by anyone they recognized as a peer. A vendor demo proves the tool works for the vendor. A colleague at the next desk proves it works for someone with their job and their constraints.

What most “adoption” buys instead

The reason this gets misdirected is that the visible moves all look like adoption work. Mandates from the top. Training modules. License counts on a dashboard. They look like effort, they look like investment, and they produce numbers. They mostly don’t produce users.

The work is unglamorous. It looks like a leader using AI in front of the team rather than telling them to. It looks like a colleague walking someone else through a real task they recognize. It looks like a hesitant person having a starting point instead of a blank chat box. And it looks like a manager saying “here’s how I’m using it” rather than forwarding an AI policy. Whether the champion is someone you hire in or someone already on the team matters less than whether they’re willing to do that work.

The six aren’t the problem; the missing conditions are. When any of those three start showing up on the team, the six tend to come around quickly.

Sometimes faster than the original two did.