World Model Readiness
Engraved role-check instrument

For you

Role Checks

Module · The gap between your talk and your use

The Executive's AI Check

You do not have to build AI, but you sign off on it, fund it, and answer for it. That takes a specific kind of judgment: enough hands-on feel to tell a real capability from a pitch, and enough discipline to know which calls a model should never make alone. This module checks the gap between how much you talk about AI and how well you actually understand it, because that gap is where expensive mistakes get approved.

Question 1 of 5 · You use it yourself

Do you use AI tools hands-on yourself, or only talk about them?

You cannot judge what you have never touched. Leaders who have actually used the tools ask sharper questions and get sold fewer fantasies, because they know from their own hands where the capability ends.

Question 2 of 5 · You ask sharp questions

When your teams bring you AI plans, do you ask questions that expose the weak spots?

The questions a leader asks set what the organisation optimises for. If you ask for the upside and the demo, you get theatre; if you ask about failure modes, data, and what happens when it is wrong, you get honesty.

Question 3 of 5 · You resist the hype

Can you tell a real AI capability from a vendor promise?

The market runs on confident claims, and a lot of them do not survive contact with your actual data and workflow. The ability to tell a demo from a deployment is what keeps you from buying the same disappointment your peers just did.

Question 4 of 5 · You know the boundary

Do you have a clear instinct for which decisions AI should never make alone?

Some calls carry consequences a model cannot own: legal, ethical, human, reputational. Knowing where that line sits, and being able to say why, is a core part of the job that no tool will draw for you.

Question 5 of 5 · You learn from failures

When AI fails inside your own organisation, do you dig into why?

The most useful signal you have is your own failures, and it is the one most likely to get buried to protect a narrative. Whether AI mistakes get examined or quietly disappeared decides whether your organisation actually learns or just repeats.

For the statistics · one click each

Three questions for the public picture

These do not affect your score. They feed the anonymised, aggregated statistics; groups under 8 respondents are never shown.

How often do you personally use AI tools for your own work?

Never
Rarely
Weekly
Daily
Prefer not to say

What do most of your AI decisions rest on?

Vendor claims
What peers do
Trusted advisors
Evidence and trials
Mostly gut feel

When an AI initiative underdelivers, what usually happens?

It gets buried
Quietly dropped
Noted informally
Formally reviewed
None yet

Your context

Used to calibrate the report. Company size and sector remain in the anonymized dataset; your email does not.

What the five levels look like

Every dimension in this assessment is scored 1 to 5. This is what the levels mean, dimension by dimension. The graded report diagnoses where your own answers land and what to do about it.

You use it yourself

  1. 1Never touch it
  2. 2Watched a demo
  3. 3Dabbled once
  4. 4Use it weekly
  5. 5Use it daily

At the low end: Approving AI you have never used is like approving a factory you have never walked. Spend a few hours a week actually working with the tools, so your instinct is built on contact, not slides. What good looks like: Hands-on daily use is what makes your judgment worth trusting. Keep using the current tools, not the ones you learned last year; the capability moves and stale intuition misleads.

You ask sharp questions

  1. 1Rubber-stamp it
  2. 2Ask for hype
  3. 3Generic questions
  4. 4Probe the risks
  5. 5Cut to the core

At the low end: A rubber stamp teaches your teams to bring you polish instead of truth. Start asking what breaks this, what data it needs, and what happens when it is confidently wrong. What good looks like: Questions that cut to the core are how you lead on AI without building it. Keep asking them consistently; the day you stop, the theatre comes back.

You resist the hype

  1. 1Buy the pitch
  2. 2Easily impressed
  3. 3Some skepticism
  4. 4Test the claims
  5. 5Hard to fool

At the low end: Buying the pitch is how budgets get spent on capability that does not exist yet. Before any commitment, insist on a trial against your own data and your own edge cases, not the vendor's demo. What good looks like: Being hard to fool is a rare and valuable trait in a buyer. Keep testing claims yourself; the vendors get better at the pitch every quarter, and so must your filter.

You know the boundary

  1. 1No line drawn
  2. 2Fuzzy sense
  3. 3Rough boundary
  4. 4Clear boundary
  5. 5Boundary and reasons

At the low end: With no line drawn, the boundary gets set by whoever deploys the tool, not by you. Name the decision types in your remit that must always keep a human accountable, and say so out loud. What good looks like: A clear boundary you can explain is exactly what your role owes the organisation. Revisit it as the tools improve; the line moves, but it should move deliberately, not by drift.

You learn from failures

  1. 1Never hear of them
  2. 2Bury them quietly
  3. 3Note and move on
  4. 4Review the causes
  5. 5Turn them into lessons

At the low end: If AI failures never reach you, someone is filtering them, and you are flying on a rosy picture. Make it safe and expected to surface what went wrong, then actually look. What good looks like: Turning failures into shared lessons is how an organisation compounds its judgment. Keep the reviews blameless; the moment they start assigning blame, the honest reporting dries up.