Make your next step concrete.

Choose one task, try a small change, then check whether it helped. You can fill in the review after your experiment.

This page keeps your entries in your browser. Nothing you type here is sent to our server or saved by this page. No AI runs here. Copy or print to keep your work before leaving.

Use generic task names and public or made-up information. Leave out names, customer data and confidential work. Follow your workplace policy when you try the experiment.

1. Plan a small test.

Pick one low-risk task you can check yourself. A specific time and a clear quality bar make it easier to try.

Task to test

When I will try it

Time budget (minutes)

Include learning and checking.

What a good result must pass

What must be correct? What must a person still decide?

2. Record what happened.

Complete this after the test. Compare the same task and quality standard. Leave times blank if you have not measured them.

Usual time (minutes)

AI attempt time (minutes)

Checking and correction time (minutes)

What worked, failed or needed judgment

3. Choose what comes next.

Repeat, adjust or stop. A test that reveals a poor fit is useful evidence too.

My next step

Add a task or load the fictional example to start. Empty fields stay out of your copy.

1. Plan a small test.

2. Record what happened.

3. Choose what comes next.

My task experiment