Free AI capability challenge
Are you as good with AI as you think?
Rate your confidence, then find out. Seven quick questions on three skills every AI user needs: checking its work, briefing it and knowing what to hand over. About five minutes, on your phone or computer.
Three skills
- 1Review. Spot the mistake. AI errors look plausible, so you'll see things AI wrote and decide whether anything is wrong.
- 2Brief. Get the prompt right. Build a prompt, run it and see how the output changes.
- 3Delegate. Know what to hand over. Decide which everyday tasks are fine for AI, which need careful checking and which aren't for AI at all.
What you get
A level from 1 to 5, a score for each skill, how your confidence stacks up against your result, and tips on what to work on. You'll see what you got right and why after every question.
Then share your level on LinkedIn, or send the challenge to a colleague and see if they can beat you.
Your result is emailed to you, never to your employer.
How it works
- 01
Tell us a little about you
Your name, email and a few quick questions about your work. No account or password.
- 02
Rate your confidence
Say how good you think you are at checking AI's work, before you see a single question.
- 03
Answer seven questions
Everyday situations, from a team update to a social post. Tap to answer. There's no timer.
- 04
See your result
Your level, your score on each skill, and whether your confidence matched how you did.
Why these three skills
It's easy to feel confident with AI. What matters is whether the work that leaves your desk is right.
Reviewing matters most because AI mistakes rarely look like mistakes. A wrong figure, a claim nobody asked for or a quote applied to the wrong people reads just as smoothly as the rest. The person who signs it off is the last check. Our guide on how to check AI output before it reaches a client covers what to look for.
Briefing decides what you get back. Saying what to work from, what to leave out and what shape you want fixes most problems before they start.
Delegating is knowing when not to use AI, or when to use it and check every line. Some tasks are low risk. Others put confidential information or client commitments on the line.
None of this needs years of AI experience. The challenge tests what good work requires, so someone who rarely uses AI can still score well.
Questions about the challenge
Is it really free?
Yes. There's nothing to pay and nothing to install. You need an email address so we can send you your result.
How long does it take?
About five minutes, including a few questions about you. There's no timer, and it works on a phone.
Who is it for?
Anyone aged 18 or over who uses AI at work, or expects to. You take it as an individual, not through your employer.
Do I need to be an experienced AI user?
No. The questions test the judgement good work needs, not how often you use AI tools.
Will my employer see my result?
No. Your result is sent to you and never shared with your employer. We only use results in aggregate, with names, email addresses and firm names removed. The privacy notice has the details.
Can I take it again?
Yes, as often as you like. Please take it on your own, without help from other people or AI tools. The challenge terms set out the rules.
Is this the same as an Osava assessment?
No. The challenge is a snapshot from seven questions. Firms that use Osava test their people on real work from their sector, reviewed by their own senior people, and track what AI is returning over time. See how Osava works.
For firms: about Osava
We show firms what their AI investment is returning, and how to get more from it. We test your people on real work, find the workflow changes worth making, and track what each one returns.