How Candid™ works

What Candid™ measures

Candid™ looks at how you work with AI: whether you check what it gives you, own the result and know when the work is beyond what you can judge. It gives two scores.

How the test works

1. You rate yourself.You say how good you think you are at the things Candid™ looks at.
2. You face everyday work situations.For each one, you choose what you would most likely do.
3. You describe your habits.Short questions on how you usually work and how you use AI.

Why situations?

People are poor judges of their own skill. What you would do in a situation shows more than a rating you give yourself, so Candid™ compares the two. Where they differ is often the most useful finding.

If a situation doesn’t fit your work

You can say so. It is left out of your score and never counts against you.

If you are new to AI

You take a shorter test about how you work now, with no questions that assume you use AI. You get a Rookie result, which sits before the nine profiles. Take the full test once you have started, to find your place among them.

Your score

Your answers give you two scores: human judgment and AI skill.

Your Candid™ score

Your Candid™ score brings the two together. It rewards AI skill you can check: work you can’t check is worth less.

Your profile

Your two scores place you in one of nine profiles.

What a score is for

A score shows how you work with AI today. It is not a verdict, and it will move as your habits change.

What Candid™ is built on

Candid™ began in our work with organisations: designing ways of working with AI, testing them and seeing where they held up and where they broke. We then checked it against published research and piloted early versions of the test. Researchers broadly agree that being good with AI takes more than knowing the tools. Candid™ puts the most weight on judging what AI produces.

What Candid™ can’t tell you yet

Candid™ is new, and some parts are still provisional.

• The lines between the profiles.They are our starting points, and we will refine them as more people take Candid™.
• How steady results are.We are still checking how results hold over time and how they relate to the way people work.
• What you say and what you do.Your answers show what you say you would do. Under real pressure, people sometimes act differently.
• Today’s tools.What counts as good AI use is based on the tools as they are today. It will change as the tools change.

Candid™ is a development tool. It is not built for hiring or appraisal.

We will update this page as we learn more.