Nontechnical employees can learn to use AI well, because using AI well at work is mostly judgment, not technology. They need to know your rules, a few plain words, which jobs to hand to AI, how to say what they want, what to show it and how to check what comes back. They don’t need code, math or how models are built. They need short, plain-language practice on their own work.
What people need first
- The rules. Which AI tools are allowed at work, and what must never go into them. A nervous person needs this more than a confident one.
- A few words. A hallucination is AI stating something false as if it were true. The context window is everything the AI can see while it answers. These two words explain most of AI’s surprises. See AI terms explained.
- Judgment. Which jobs to hand to AI, which to check first, and which need a different tool. This habit prevents the most harm, and it uses knowledge people already have about their own work.
After that come briefing (saying what you want back, the limits and what counts as done), context (what to show AI and what to keep out) and checking (reading its answer against the facts).
What they don’t need to know
- How to code. Not even to make a reusable skill: a SKILL.md file is plain instructions in a text file.
- How models are built. Training, parameters and the math behind them are interesting, but they don’t change what a person should hand to AI on Monday morning.
- Prompt tricks. Magic phrases come and go. A clear brief, the kind you would give a new colleague, works better.
- Every tool on the market. Getting good at the one assistant your company allows is worth more than a tour of ten.
Their existing expertise is the advantage
The hardest part of using AI well is knowing when its answer is wrong. That takes knowing the work. An accountant spots the figure that doesn’t match the ledger. A project manager spots the date that moved. Someone who has done a job for years is often better at checking AI on it than a technical colleague who hasn’t. Tell people that, because many assume the opposite.
Practice in plain language
Good practice for nontechnical people looks like their own week, not a computer science class:
- Situations from their role. A client follow-up, a meeting summary, a month-end note. Not a coding exercise.
- A decision, not a lecture. “Which of these would you hand to AI?” rather than a video about AI.
- Feedback right away, with the reason in plain words.
- Mistakes come back for another try, until they get it.
- Short and regular. Practice that fits between other work beats a long session once a quarter.
Avoid streaks, leaderboards and countdowns. They put off exactly the people who are least sure of themselves.
Common fears, and honest answers
“I’ll break something or leak something.”
This is the most useful fear, so answer it with specifics: which tools are approved and what stays out of them. Then practice the call itself, deciding what a document holds before anyone pastes it in. A clear rule turns worry into a habit.
“I’ll look stupid.”
Let people practice where a wrong answer is private and simply comes back for another try. If you use a training product, check what managers can see, and tell people before they start.
“I’m not a tech person.”
Most of what matters is judgment about their own work, which they already have. The vocabulary is small and learned by using it.
“AI gets things wrong, so why bother?”
It does get things wrong. That is why the habit of checking matters, and why the people who know the work are the right people to use it. Start with jobs where a mistake is easy to catch, such as a first draft someone will read anyway.
“Is this going to replace me?”
Don’t make promises nobody can keep. Say plainly how your company plans to use AI, and what you are asking people to do with it. Training that builds their own judgment is an investment in them, and you can say so.
How to start a nontechnical team
- Publish the rules in one page, in plain words.
- Pick one job everyone on the team does, and practice handing it to AI together.
- Give people short, regular practice set in their own work.
- Ask after a month: what did you do differently in your last chat with AI?
For the wider plan, see how a company should train employees on AI. Skill Tree’s first lesson plays in the browser with no account, so anyone can try one before you decide.