Skip to content

Pick a problem worth an hour.

The first project decides whether you keep going. Choose one you will finish and can show.

Small, real and yours

People stall on a first AI project because they choose something too big, too abstract, or too far from their own work. A good first project is small enough to finish in an hour and real enough that you care about the result.

Five steps

  1. 01

    Choose a weekly task

    Something you do again and again, that is annoying but not risky: a summary, a reply, a report.

  2. 02

    Write what good looks like

    Three sentences. Without this you cannot tell whether the AI helped.

  3. 03

    Try one real example

    Use one real input, with private details removed. Do not start with a toy.

  4. 04

    Check the result

    Use the five checks from the guide on checking an answer. Fix what is off.

  5. 05

    Write it up

    Half a page: the task, what you tried, what worked, what you would change. This is your first proof of work.

How to know it is the right size

If you cannot describe the finished result in one sentence, it is too big. Cut it until you can.

  • It uses one input and produces one output
  • You can tell if it worked in a minute
  • A mistake would be annoying, not harmful

Take the idea to a Labs project.

Start with a free project in Labs, or build your first AI solutions in the free Masterclass.

  • Learn by building, from the first session
  • Free to start
Get early access

Take the idea to a Labs project.

The first project decides whether you keep going. Choose one you will finish and can show.