How to Train Your Team on AI in One Hour

Most AI policies sit unread in a shared drive. If you wrote one, filed it, and moved on, your team is now using AI tools however they feel like it. Some are sharing client data with chatbots. Some are refusing to use AI at all. Most are doing something in between, without any real guidance.

This article gives you a session plan you can run this Friday. One hour. No slides required. No specialist knowledge needed. Just a room, a laptop, and the policy you already have.

What This Session Is For

The goal is not to teach AI. The goal is to make the policy real. People follow rules they understand and have seen in practice. They ignore rules they have only read, or never read at all.

By the end of this hour, your team will know three things: what they are allowed to do, what they must not do, and why those limits exist. That is enough to change behaviour.

The Full Agenda

Block by Block: What to Do and Say

0:00 to 0:05, Why We Are Here

Open with one sentence: "We adopted AI tools a while ago and I want to make sure everyone knows what the rules are and why."

Do not apologise for calling the meeting. Do not oversell AI. Do not say this will change everything. Just be direct.

Tell them you are going to show them some real mistakes, give them time to try a prompt themselves, and hand them a single page to keep. That is the whole session.

0:05 to 0:15, What the Policy Actually Says

Read the policy out loud, together. Do not email it beforehand. Do not assume people have read it. Read it in the room.

Stop after each section and ask: "Does that make sense? Any part of that unclear?"

Highlight three specific things verbally as you go:

1. Which data categories must never go into an AI tool (client names, financial records, personal data, anything covered by an NDA).
2. Which tools are approved and which are not.
3. Who to contact if someone is unsure.

If your policy does not cover those three things clearly, note it as a gap to fix after the session.

0:15 to 0:30, Live Demonstrations: Three Mistakes to Show

This is the most important block. Do not describe mistakes. Show them.

Open a free AI chat tool on your laptop, projected or shared on screen.

Mistake one: pasting in real client data.

Type something like this: "Here is a summary of our meeting with [Client Name] on [date]. Their revenue is [figure]. Please write a follow-up email."

Stop before you press enter. Say: "This is the mistake. That client name, that revenue figure, that date. Those details are now going into a system we do not control. If our policy says client data must not be shared with external AI tools, this breaks the rule."

Delete it. Show them the empty box. "That is what we send instead: no names, no figures, no identifying details."

Mistake two: accepting AI output as fact.

Ask the tool a specific factual question relevant to your industry, something with a clear right answer. Let it respond. Then point out anything it got wrong or oversimplified. If it got everything right, ask a follow-up that pushes it into guesswork. It will usually oblige.

Say: "AI tools produce plausible text. Plausible is not the same as accurate. You still check the output."

Mistake three: using a non-approved tool.

Show them a browser search for "free AI writing tool" and click the first result. Do not type anything. Just show them the homepage.

Say: "This tool has different privacy terms from the ones we reviewed. We do not know what it does with input. That is why the policy lists approved tools. If a tool is not on the list, ask before using it."

Three mistakes, shown, not described. Takes about twelve minutes if you move at a steady pace.

0:30 to 0:45, Hands-On: One Prompt Each

Ask everyone to open an approved AI tool on their own device. Give them one task: write a prompt for something they actually do in their job, following the data rules you just demonstrated.

Walk around the room. If someone's prompt includes real names or sensitive details, stop them and rework it with them, out loud, so others can hear.

If someone writes a genuinely good prompt, ask them to share it with the group. This builds confidence without turning the session into a lecture.

0:45 to 0:55, Questions and Common Scenarios

This is where two characters usually appear. Handle them directly.

The sceptic. They say: "I don't really see the point, I've been doing my job fine without this."

Respond without arguing. "That is fair. The policy is not telling you to use AI. It is telling you what the rules are if you do. You are welcome to not use it. Just know that colleagues might, and this session is for everyone."

Do not try to convert the sceptic. They often become your best risk managers if you treat their caution with respect.

The over-enthusiast. They say: "Can we use it for X, Y, Z? What about automating the whole onboarding process? I read that you can connect it to our CRM and..."

Respond specifically. "Good question. That kind of integration is outside what we have approved so far. Add it to the list and we will review it properly before anyone sets it up."

Give the over-enthusiast a role: ask them to bring two specific tool suggestions to the next team meeting, with a short note on what each one does and what data it would touch. That channels the energy without letting it run ahead of your governance.

What to Hand Out at the End

At 0:55, hand every person one sheet. It should cover: approved tools, data rules in plain language, who to ask if unsure, and the date of your next review.

That is the session. Sixty minutes. No slides. Real demonstrations. One page to keep.

What Comes Next

Running this session once is not enough. You need a 90-day plan that schedules two more check-ins, a way to track whether AI use is saving time or creating problems, and a library of approved prompts so people are not inventing their own from scratch.

The Plainstart AI Adoption Kit contains all of that. It includes the staff one-pager designed to be the handout from this session, the 90-day adoption plan with built-in review points, a prompt library for common business tasks, and a tool scorecard for evaluating anything your enthusiast brings to the table. The full kit is $149. The AI Usage Policy that anchors everything is free.

This is general business guidance, not professional, legal, or compliance advice. If your industry has specific regulatory obligations around AI use, check with a qualified adviser before finalising your policy.

Plainstart. AI adoption, done properly.