Back to news

OpenAI maps ChatGPT work use: where UK small firms could start

Share
OpenAI maps ChatGPT work use: where UK small firms could start

AI Data Study

OpenAI has mapped 165 work tasks people hand to ChatGPT across nearly 100 countries — UK small firms will get more from it by picking one task than by reading the whole list.

165
Work activities in OpenAI's ChatGPT usage map
Drawn from OpenAI Signals data across ChatGPT users worldwide
100
Countries covered in the OpenAI Signals dataset
Nearly 100 countries; does not isolate UK business use specifically

What OpenAI has published

OpenAI has released ChatGPT's expanding world of work, a look at its OpenAI Signals data. It sorts how people use ChatGPT for their jobs into 165 work activities, across nearly 100 countries, and tracks how that use changes over time.

This is a useful map of what people already do with the tool. It shows writing, research, planning and analysis tasks across many kinds of work. For a business owner wondering where AI might fit, it is a menu of real-world uses.

The data has limits. It comes from OpenAI's own users around the world, so it tells us how ChatGPT users behave in general. It does not measure how UK businesses use AI, and it cannot tell you which tasks will work well in your firm. Treat it as a list of ideas to test.

Why a long list can slow small firms down

A map of 165 activities is exciting, and it can also be a lot to take in. Small teams rarely have spare hours to trial AI across the whole business at once. Spreading a first attempt too thin often means nothing gets tested properly, and the tool gets dropped before anyone sees a result.

A better route is to choose a single job and give it a fair trial. The best first task has three traits:

  • It eats time. Something your team does often, such as first drafts of product descriptions, summaries of long supplier emails or tidying meeting notes.
  • It carries low risk. If the output is weak, nobody outside the business is harmed, and nothing goes out without review.
  • A person checks it. Someone who knows the subject reads every output before it is used.

Skills and trust matter as much as the tool. Staff need time to learn how to ask for what they want, and they need to see that the results hold up before they rely on them. A small, contained trial builds both. Once one task works, the team has a template it can copy for the next one.

What this means for UK businesses

  • Use OpenAI's data for ideas. It shows what people around the world do with ChatGPT. Your own trial will tell you what works for you.
  • Start with one task. Choose work that is frequent, time-heavy and low-risk, and resist trying everything at once.
  • Keep a human check. Every output should be read by someone who knows the subject before it goes anywhere.
  • Build skills as you go. Give the people running the trial time to learn and share what works.
  • Decide what success looks like first. Time saved and quality kept are simple measures any team can track.

Analysis compiled by Sayu (sayu.co.uk), a UK digital marketing agency.

Share