People09 Sep 2026

Five skills for getting good results from AI

By Alan Goh · 2 min read

One loop

Five skills to direct AI

  • Break it down
  • Give it the context
  • Set the limits
  • Refine it
  • Judge the result
Short answer

Getting good results from AI comes down to five skills: break the task down, give it the context, set the limits, refine the draft, and judge the result. When an answer disappoints, the missing skill tells you what to fix.

Two colleagues use the same AI tool for the same kind of task. One has something usable in a few minutes. The other spends half an hour rewording the request and gives up.

What separates them is a small set of skills, and every one of them can be learnt.

The five

1 · Break it down. Large requests get vague answers. Split the job into steps and ask for one at a time: the outline, then each section, then the summary. When AI seems to be guessing, the task usually needs breaking down further.

2 · Give it the context. Who it’s for, what it’s for, what has happened so far, when and where it applies, and how it will be used. The background a capable colleague would want before starting.

3 · Set the limits. Length, format, tone, and anything to leave out. Limits narrow the many possible answers down to one you can use.

4 · Refine it. Treat the first answer as a draft and steer it: shorter, warmer, more detail on the second point. Each round gets closer.

5 · Judge the result. Read it the way your reader will. Is it right? Does it do the job? Would you send it?

Why it’s a loop

The five work together, and when a result comes back wrong, the loop tells you which skill to reach for.

Muddled or rambling? Break it down. Generic? Give it more context. Too long, or in the wrong format or tone? Set the limits. Close, with something still missing? Refine it. Unsure whether it’s correct? That is the fifth skill at work, so check before you use it.

A disappointing answer becomes a diagnosis. You go back to the skill that was missing, adjust, and run it again.

They travel with you

These skills belong to the person, and they work with any AI tool your team picks up. They are also the skills you’d use to brief a capable new colleague, which is why they feel familiar once you start.

Practice makes them stick

Each skill clicks fastest through practice on everyday tasks. In my sessions we practise each one on something simple first, such as explaining how to tie a shoelace, so the method is clear before the real work begins.

The test

Next time an AI answer disappoints, name the skill that was missing. Use that one, and run it again.

In your department

  • HR & Admin: When you plan AI training, these five skills are what to look for afterwards. They apply in every department and with every tool.
  • Sales & Marketing: A generic campaign email usually means missing context: who the customer is, what the offer is and why they should act now.
  • Strategy & Planning: Break a large planning question into steps, take them one at a time, and judge each answer the way your board will read it.

Want to try this with your team?

Tell me what your team is working on, and I'll suggest where to start. I read and reply to every message myself.

More notes

  • 02 Sep 2026

    Keep your AI instructions in a file

    The instructions you give AI for a repeated task are worth keeping in a plain document. Use the file every time, update it whenever you correct a result, and the task gets better with every run, whoever on the team runs it.

    Read it →
  • 19 Aug 2026

    Split AI work into asking, doing and checking

    AI work goes better with three roles kept apart: you decide and write down what good looks like, a fresh AI session does the work from that instruction alone, and someone with no stake in the result checks it against your description.

    Read it →
  • 05 Aug 2026

    AI can be confidently wrong

    AI writes by predicting what fits, so it can give an answer that is specific, fluent and wrong. Treat every AI answer as a draft, and check the parts that carry weight, such as numbers, references and names, before you rely on it.

    Read it →

All the notes →