Your Team Isn't Adopting AI. They're Just Pretending To — Here's Why It Matters

A workplace where everyone feels valued is the first step to building successful AI integration. (Image by Mesaros-Winckles generated by Google Gemini)

When I talk with business owners and executives about their AI rollout, I hear a version of the same story: "We've adopted AI." What that usually means is everyone got a license — ChatGPT, Gemini, Claude — and a few people use it to answer questions or brainstorm. That's access. It is not adoption.

Real adoption means your team understands how the tool actually helps them, not just that they're allowed to open it. So when I start working with a team, I ask one question that tells me more than any usage report ever could: Do you use an AI tool to help balance your own strengths and weaknesses? Answering that honestly takes real self-awareness — and it takes psychological safety, because you have to admit where you're weak before a tool, or anyone else, can help.

That's part of why I built what I call a safe chat into my workshops: a private Gemini conversation where people can ask real questions about AI adoption and their own comfort level, with no name attached to their answers, which I then fold into one anonymized report for the organization. It surfaces the same thing almost every time — questions people are too nervous to ask out loud. They're afraid of looking clueless in front of the team, or of admitting they've been using AI in a way their boss hasn't sanctioned. That fear is psychological safety showing up in real time, or the lack of it.

So here's the question worth asking before you buy one more license, book one more training, or send one more memo about "embracing AI": Is your team actually using it, or just pretending to?

For small and midsize employers, that gap between access and adoption is rarely about the technology. It's about whether people feel safe enough to say, out loud, "I don't know how to do this yet." That's psychological safety, and it's what determines whether an AI rollout sticks — not the tool you picked, not the training deck, not the one early adopter who won't stop talking about it.

What Psychological Safety Actually Means

Harvard Business School researcher Amy Edmondson coined the term in 1999, long before anyone worried about a chatbot sitting in on the weekly status meeting. Her definition still holds: psychological safety is a team's shared belief that it's safe to ask a question, admit a mistake or push back on an idea without getting punished for it. It isn't "everyone gets along." A safe team can disagree and argue — it just can't punish someone for saying "I don't get this."

In fact, Edmundson's research is even more relevant now in the age of AI. This past February, Edmundson told the HBR "On Strategy" Podcast that AI creates "predictable patterns of team dysfunction" that mirror classic organizational problems. AI doesn't invent trust problems. It exposes the ones already there.

Why AI Raises the Stakes

AI stacks new fears on top of ordinary team friction: looking replaceable, if the tool can draft what you draft. Looking incompetent next to something that sounds confident even when it's wrong. Both push people toward the same shortcut — fake it, nod along, copy the output and hope nobody asks how you got there.

Workplace psychologist Paula Davis reported in Forbes in May that a striking share of workers already treat AI like a coworker, and some trust it more than the humans they work with. That's not a technology problem. It's a trust problem, and better software won't fix it.

Three Habits That Build Safety During a Rollout

None of this requires an HR consultant. It takes three small, repeatable habits.

Say your own learning curve out loud. If you're still fumbling with the tool, tell your team. "I had to ask it three different ways" builds more trust than any rollout memo.

Put the trial period in writing.

1. Don't just say "we're testing this" — write down that mistakes are expected and won't be held against anyone. A spoken reassurance evaporates; a written one is still there three weeks later.

2. Ask a better question than "Is this working?"

3. Try "Where is this getting it wrong for you?"

   
   

The first invites a polite yes. The second assumes friction exists and gives people permission to name it.

What This Looks Like for a Small Team

Scale it down. A five-person team can run a five-minute Friday check-in: What did the AI tool get right, get wrong, and what did you end up doing by hand instead? Keep a running list of wins and misses on a whiteboard or shared doc — no survey software required. The goal is making it normal, weekly and low-drama to say "this didn't work for me."

Where to Start

If you suspect your team has access instead of adoption, that's worth a closer look before you buy one more license. It's why I built the safe chat into my workshops — a private, anonymized way to find out what your team doesn't understand yet. Prompt & Proper's AI Readiness & Workflow Assessment is a low-pressure way to start that process, and building these habits into a rollout plan from day one is core to the workshops and consulting sessions Prompt & Proper runs for small teams.

The technology is supposed to make communication easier, not replace it. Get the trust part right, and the tool part tends to follow.