World Model Readiness
Engraved organisation instrument

For your company

Organisation & People

Module · Who builds it, who keeps it running

The AI Talent & Skills Assessment

Every AI roadmap assumes people who can build, integrate, and maintain it. This module tests whether that assumption holds: the depth of your bench, the realism of your hiring plans, how much you lean on consultants who leave with the knowledge, and whether the people who understand your systems have any reason to stay.

Question 1 of 5 · Bench can build it

Can your own people build and run the AI systems on your roadmap, without a vendor in the room?

Not whether anyone has touched a model: whether you have people who can ship an AI system into production and keep it alive when it breaks at 3am. Slideware skills and delivery skills are different things.

Question 2 of 5 · Hiring plan is real

Is your plan to hire the AI skills you need grounded in what the market will actually give you?

Job descriptions that ask for five years of experience in a two-year-old field do not get filled. The question is whether your hiring plan matches salary bands, timelines, and the competition you are up against.

Question 3 of 5 · Upskilling actually happens

Are you deliberately turning existing staff into people who can work with AI?

The cheapest AI talent already works for you and knows your business. Without a real programme, budget, and protected time, upskilling stays a line in a strategy deck that nobody acts on.

Question 4 of 5 · Key people will stay

Do the people who understand your AI systems have a reason to stay?

The person who built the model carries context no documentation holds. If they can double their salary elsewhere and nothing but inertia keeps them, you are one resignation away from a black box.

Question 5 of 5 · Not consultant-dependent

If your consultants walked off the project tomorrow, could you keep the AI systems running?

Consultants are fine for a push and dangerous as a permanent crutch. The test is whether knowledge transfers to your staff as you go, or whether every system they build leaves with them.

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.

Where does your AI expertise mainly come from today?

Mostly in-house
Mostly consultants
Roughly even mix
Almost none yet

How hard is it to hire the AI talent you need?

Not hard
Manageable
Very hard
Effectively impossible
Not hiring for it

If your lead AI person resigned tomorrow, what happens?

We would be fine
Slowed but coping
Serious disruption
Systems at risk
We are not sure

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.

Bench can build it

  1. 1No in-house capability
  2. 2One or two people
  3. 3Small stretched team
  4. 4Solid core team
  5. 5Deep, redundant bench

At the low end: A roadmap with no bench behind it is a wishlist. Before approving the next phase, name the people who will build and run it; if you cannot, the plan is really a hiring plan in disguise. What good looks like: A deep, redundant bench is your rarest asset. Protect it: keep the work interesting and the knowledge shared, because this is what competitors cannot buy off the shelf.

Hiring plan is real

  1. 1No plan
  2. 2Wishlist, no budget
  3. 3Budgeted, unrealistic
  4. 4Realistic but slow
  5. 5Funded and filling

At the low end: Hiring with no plan means you compete for scarce talent by accident. Write one grounded in real salary bands and timelines before you post another role that will not get filled. What good looks like: A funded plan that is filling is working. Keep the pipeline warm even when you are not hiring; the best candidates appear on their schedule, not yours.

Upskilling actually happens

  1. 1Nothing organised
  2. 2Occasional courses
  3. 3Voluntary, unfunded
  4. 4Funded programme
  5. 5Structured, time-protected

At the low end: Your cheapest AI talent already works for you. Start one funded, time-protected programme this quarter; unfunded good intentions have never trained anyone. What good looks like: A structured, time-protected programme compounds: this year's learners mentor next year's. Tie it to real projects so the skills land in production, not just in certificates.

Key people will stay

  1. 1Actively at risk
  2. 2Nothing holding them
  3. 3Market-rate, no more
  4. 4Retained and engaged
  5. 5Locked in, invested

At the low end: People who understand your systems and have no reason to stay are a resignation waiting to happen. Find out what would keep them before the counter-offer, not after. What good looks like: Retained, invested people are your continuity. Keep documenting alongside them anyway; the goal is a team that would survive even the departure you are not expecting.

Not consultant-dependent

  1. 1Totally dependent
  2. 2They hold everything
  3. 3Some handover
  4. 4Knowledge transferring
  5. 5Self-sufficient

At the low end: Total dependence on consultants means you are renting a capability you will never own. Insist on knowledge transfer as a contract term now, before the systems they build become ones only they understand. What good looks like: Self-sufficiency after consultants leave is the right end state. Keep it by doing the next build in-house with consultants advising, not the reverse.