How to learn AI when you are not technical

Many people want to learn AI but have no idea where to start. They assume you need to code, or understand the maths underneath. You don't. To use AI well at work, what you mostly need is a method and a little regularity.

Start from your tasks, not from the tool

The first mistake is starting with the tool. You open an assistant, type a random question, get an average answer, and close it. AI stays a curiosity, not a help.

Do the opposite. Take a sheet of paper and write down what you do in a normal week: emails, meeting summaries, presentations, reviewing documents, finding an idea for a post. Then circle two or three tasks that come back often and take you time.

These are your first training grounds. You are not learning AI in general. You are learning to do things you already do, better and faster.

Pick one tool and practise on real work

There are many tools, and there is no need to test them all. Choose one general assistant, such as ChatGPT, Claude or Gemini, and stick with it for a few weeks. You progress faster by knowing one tool well than by jumping between several.

Then practise on your real tasks, not on made-up examples. If you have to write meeting notes tomorrow, write them with AI. You will see right away what works and what does not, in your own context.

Learn to give clear instructions

This is the heart of it. AI answers what you ask. If the request is vague, so is the answer. Think of a very fast new colleague who knows nothing about your company or your job.

A good request usually has four parts:

1. Context: who you are, the situation, who the result is for.
2. Role: for example "act as an HR manager" or "act as a patient teacher".
3. Examples: an email you already wrote, a tone you like, a model to follow.
4. Format: a table, five bullet points, ten lines, a formal or friendly tone.

Compare "write an email to a client" with "I run a small agency, a client has been waiting a week for a quote, write a short, polite email apologising for the delay and offering a call on Thursday". The second request gives a result that is almost ready to send.

If the result is not right, do not start over. Say what is wrong and ask for a correction. It is a conversation, not a one-off order.

Always check, and protect your data

AI makes mistakes. It can invent a number, a date or a source with total confidence. You stay responsible for what you send.

Read every answer. Check important facts against a reliable source. Before sending, ask yourself one simple question: would I sign this?

The second rule matters just as much. Do not paste confidential information into a tool without knowing what it does with it: client data, salaries, contracts, health information. Check your company's rules and, when in doubt, replace names and numbers with examples.

A simple 30-day plan to learn AI

The secret is regularity, not intensity. Fifteen to twenty minutes a day for a few weeks beats one full day, once.

• Week 1, discover. Pick your tool. Every day, use it for one small real task: rewording an email, summarising a text, preparing a list.
• Week 2, ask better. Apply the four parts: context, role, examples, format. Compare with your results from week one.
• Week 3, check and correct. Get into the habit of reviewing and asking for improvements. Note the mistakes AI often makes in your field.
• Week 4, build your templates. Keep your best requests in a document and reuse them. Share one with a colleague.

At the end of the month, you will not be an expert. But you will have a habit, a few requests that work, and a clear idea of what AI can do for you.

Learn with others, then go further

People learn better together. Ask two or three colleagues to share, every week, one use that saved them time, or one mistake they spotted. Other people's mistakes save a lot of time.

Once the habit is in place, structured training helps you take the next step: organising your work with AI, using it as a team, or applying it to a specific job. The essentials stay the same: start from your real work, and practise.