How to Get Your Team Started with AI

# How to Get Your Team Started with AI

Most teams don't fail at AI because of the tools. They fail because no one gave them a clear starting point.

I've watched it happen more than once. A company buys ChatGPT licenses for everyone. An excited email goes out. A few people play with it for a week. Then it quietly fades into the background, used by maybe two people who already liked tech.

That's not an AI problem. That's an onboarding problem.

Here's what I've found actually works when you want to get a non-technical team using AI in their real work.

Start with one job, not the whole company

The worst thing you can do is say "we're rolling out AI" and leave it at that. That framing puts the burden on each person to figure out how AI fits their specific job. Most won't bother.

Instead, pick one role or one workflow and go deep on that first.

For example: your operations manager spends two hours every Monday pulling together a status report from three different systems. That's a real pain point. That's where you start.

Once one person on your team has a genuine win — something that actually saves them time or headache — that story spreads. You don't have to sell it. They do.

Find the tasks worth automating

Not every task is worth touching with AI. The ones that are tend to share a few traits:

They're repetitive. They take longer than they should. They follow a predictable pattern. They involve summarizing, drafting, sorting, or translating information.

Things like:

• Writing the first draft of a client email based on notes
• Summarizing a long document into key points
• Turning a messy meeting transcript into action items
• Drafting a response to a common customer question
• Pulling insights from a spreadsheet without writing a formula

Notice that none of these require a non-technical person to learn how to code. You're just redirecting effort that was already happening.

Give your team a prompt that works

Here's a problem I've seen over and over: someone opens ChatGPT, types "help me write an email," gets back something generic and useless, and decides AI isn't for them.

The tool isn't the issue. The prompt is.

When I introduce a team to AI, I give them a template prompt for their most common task. Something like:

"I need to write an email to a client following up on their invoice. It's been two weeks since it was due. The amount is $4,200. The client is friendly but slow to respond. Keep it professional but warm, under 100 words."

That's specific. It tells the AI who the audience is, what the goal is, and what constraints matter. You get a useful draft on the first try.

Once someone sees that work, they start building their own prompts. But they need that first win.

Run a short working session, not a training

There's a big difference between a training and a working session.

A training is: here's a slide deck, here are the features, here's a demo, good luck.

A working session is: everyone bring one real task you've been procrastinating on. We're going to try to finish it together using AI.

Working sessions win every time. People leave having actually done something. They don't just know about AI — they've used it.

I usually run these in 60-90 minutes. No slides. Just a simple agenda: intro the tool, model one example task with the group, then everyone works on their own thing and we share what we got.

The side effect is that people teach each other. One person figures out a clever prompt and shares it. Someone else figures out a shortcut. The collective knowledge compounds.

Don't try to boil the ocean

I'll be honest: the teams that struggle with AI adoption usually have a leadership problem, not a technology problem.

Someone at the top said "we need to be using AI" without answering the follow-up questions: What specifically? Who owns it? What does success look like in 30 days?

When there's no clear owner and no clear goal, nothing moves. People wait for someone else to figure it out.

If you want your team to actually adopt AI, someone has to own it. Doesn't need to be a full-time job. But someone needs to be the person who runs the working sessions, collects the wins, and keeps the momentum going.

The practical starting point

If you're reading this and thinking "okay, but where do I actually start" — here's a simple framework:

Pick one person who's already curious. Give them a ChatGPT or Claude account and one specific task to try AI on. Check in with them after a week. If they have a win, document it. Turn it into a short example you can share with the rest of the team.

Repeat that with two or three more people. By the time you do a group session, you already have real examples from inside your own company. That's far more convincing than anything I could show you.

AI adoption doesn't require a big rollout or a consultant. It requires a clear starting point and someone willing to try.

Start small. Get one win. Build from there.