Free · four minutes · no account

What is your AQ?

AQ is your AI Quotient. A score from 1 to 10 for how capably you actually work with AI. Not what you know about it. What you catch when it gets something wrong.

Nine in ten people now use AI at work, and almost nobody has ever been told whether they are any good at it. Twenty-six questions, four minutes, and you leave knowing your level, where you rank against others in your role, and the one gap holding you a level below.

Test my AQ
AnalystsMarketersFinance OperationsProductFreelancers

For anyone who uses AI on real work. You pick your role first and the questions follow it. Nothing technical.

The ten levels

Where do you think you land?

7/ 10
Resilient
System tier
1 Newcomer10 Architect
Find out if you are right

Trusted by teams at

Catapult DigitalThe Citation GroupExperian Intent HQWistron
Why this is worth four minutes

The market has already started paying for this. Quietly.

Not for knowing about AI. For being able to use it on real work and tell when it is wrong.

+62%

Average wage premium for roles that require AI skills, up from 57% a year earlier

PwC Global AI Jobs Barometer, 2026
1 in 6

Workers who feel genuinely prepared to use AI well, out of the nine in ten who use it

Study.com, 2026
69%

Growth in roles requiring AI skills, against 9% for the job market as a whole

PwC Global AI Jobs Barometer, 2026

Almost everyone learned AI the same way: opening a chat window and guessing. No feedback, no standard, and nothing to measure against. Which is why so many people use it every day and so few have any idea whether they are good at it.

Four minutes tells you which side of that line you are on.

Test my AQ
One question from the real assessment

Everyone tests what you know. We test what you catch.

Six of the twenty-six questions put real AI output in front of you with a real mistake inside it. They are scored on accuracy and weighted double, so there is no talking your way to a high number. Here is one.

Scenario · VerificationWeighted double
AI-drafted variance commentary
Q3 operating expenditure closed at £2.41m against a budget of £2.30m, a favourable variance of £110k, driven mainly by lower headcount costs. Recommend no corrective action this quarter.

What is wrong with it?

£2.41m against £2.30m is £110k over.

✓It is well written and internally consistent
!The conclusion is inverted. An adverse variance reported as favourable, with a recommendation to do nothing about it.
This is what AQ measures. Not whether you can produce output. Whether you would have caught this before it reached a board pack with your name on it.
See the other twenty-five
What gets measured

Four dimensions, one score.

Your AQ is one number, but the useful part is which of the four is dragging it down. Most people are strong on two and have never thought about the other two.

01

Scope

How much of your actual work AI touches, and whether it handles whole processes or only isolated tasks.

Does it run, or do you drive every step?

02

Verification

Whether you can tell when the output is wrong. The dimension almost nobody scores well on, and the one that decides whether AI is safe to rely on.

Would you have caught it?

03

Quality

Whether good output is something you produce reliably or something you get lucky with after four attempts.

Repeatable, or a coin flip?

04

Transfer

Whether anything you have built is usable by someone else, or lives entirely in your own head.

Could a colleague run it?

Find my weakest dimension
At the end

Your score, then the route out of it.

  • Your AQ, 1 to 10. With the level name and what it means in practice.
  • How close your guess was. Most people are two levels out. Being wrong is the useful part.
  • Where you rank in your role. Your percentile against others doing the same job.
  • Your weakest dimension. Named, explained, and what it is costing you.
  • Your scenario accuracy. How many of the real AI mistakes you actually spotted.
  • A roadmap to level 8. Step by step, built around the job you do.
Your AQMarketing · 26 questions
5/10
Self-checking
Top 48% · you guessed 7
Weakest: Verification. You can get AI to produce good work. You cannot yet reliably tell when it is wrong, and that is what holds the score at five.
Test my AQ

Your score is free. The full roadmap needs an email.

Reasonable questions

No. Six of the twenty-six questions show you real AI output containing real errors and score you on whether you catch them. Those are weighted double, and failing them caps your score no matter how you answered the rest. You cannot self-report your way to a high number.

No. It measures how much of your work AI handles and whether you can tell when it is wrong, not whether you can run a model. Level 10 is setting the standard for a team, not writing code.

About four minutes. Twenty-six questions, most of them one tap. You pick your role first so the examples match the work you actually do.

Not to see your score. You get your AQ, your level, your percentile and your weakest dimension without giving anything. The full roadmap to level 8 is where we ask for an email.

Most working professionals land between 3 and 5. Seven and above is where employers start paying a premium. Almost nobody scores 9 or 10, and the people who do are usually setting AI policy for a whole function.

Etiq makes an AI upskilling platform, so yes, eventually. The assessment stands on its own though. Take it, get your number, and do nothing else with it if that is all you want.

You already guessed your number. Go find out.

Four minutes, free, no account. Most people are two levels off.

Test my AQ

Your score is free. The roadmap needs an email.

LinkedIn
info@etiq.ai
Contact us
©2026 Etiq AI
Privacy Policy
Terms & Conditions