A lot of businesses have already paid for AI. Copilot licences, ChatGPT accounts, the AI features built into their CRM. And in a lot of those businesses, a few keen people use it daily, and everyone else tried it once, got a mediocre answer and went back to doing things the old way.
That is not a technology problem. It is a training problem, and usually the wrong kind of training.
Why generic courses do not stick
There is no shortage of free AI courses. They explain what a large language model is, show a few clever prompts, and hand out a certificate. They are fine for curiosity. They rarely change how anyone works on Monday morning.
The reason is simple: people do not care about AI. They care about the pile of work on their desk. Training that does not start from that pile does not stick.
Start with their real tasks
The best AI training we run starts by asking each team one question: what do you do every week that is repetitive, slow or annoying?
The answers are specific. Writing quote follow-ups. Summarizing a long email chain before a call. Turning site notes into a tidy report. Pulling numbers out of a supplier PDF. Drafting a job posting.
Then we work through those exact tasks, with their own documents, in the tools they already have. By the end of the session, people have not just learned about AI. They have used it on something they would have done anyway, and saved time doing it.
Address the worry out loud
Some people are not using AI because they are quietly worried. Will this replace my job? Am I allowed to put customer information into it? What if it gets something wrong and I get blamed?
Those are fair questions, and they deserve straight answers from the owner, not silence. In our experience, three things help:
- Say why you want people to use it. "So you spend less time on admin and more on customers" lands very differently from "be more productive."
- Set clear rules. What information is fine to use, what is not, and which tools are approved.
- Make checking the norm. AI drafts, people decide. Nobody gets blamed for catching a mistake.
Make it part of the week
A one-off workshop gives people a good start. What keeps it going is small and regular:
- A short shared list of prompts that work for your business
- One person on each team who is the go-to for questions
- A few minutes in a team meeting to share something that saved time
None of this needs a big programme. It needs someone to keep it visible for a couple of months until it becomes normal.
Use what you already pay for
If you have Microsoft 365, you may already have Copilot available, or be able to add it. If you use HubSpot, it has AI features built in. Before buying anything new, get the team using what is already there. It is the cheapest and fastest win.
What good looks like
After a few weeks, you should see people reaching for AI without being told: drafting the awkward email, summarizing the meeting, checking a document before sending it. Not for everything. For the tasks where it clearly helps.
That is the goal. Not an AI-first company. Just a team that knows when this tool saves them time, and uses it.