Most organisations misjudge their own AI stage.
Adoption isn't a yes or no. It's a profile across eight stages — and the shape of that profile tells you what to do next. Drag the sliders below to see yours.
Leading edge
Stage 4
Load-bearing floor
Level 3
Overhang
None
This is what most organisations look like. Drag to match yours.
Adjust your profile
Your shape
The cliff
Executives can describe the AI strategy in detail and no one below them can describe what changed in their own job. Pilot results are quoted from a slide more than six months old.
Governance is built, tools are licensed, pilots ran and worked. Then nothing scaled, because the people layer was never built. The organisation has infrastructure and no capability.
Your shape is The cliff.
What you'll walk away with
- An evidence-based profile across all eight stages, next to the gut read you just made
- Your leading edge, load-bearing floor and overhang — three numbers that say more than any overall score
- Your responsibility gap: how far your adoption has run ahead of your responsible AI practice
- The questions we'd need to answer with you before anyone could responsibly sequence a plan
What this won't do
It won't hand you a twelve-month plan. Sequence depends on things a self-assessment can't see — who actually owns what, what's really in production, what your frontline would say if you asked them. Getting that wrong in a regulated organisation does real damage, so we don't guess at it. You'll get an honest reading of where you are and the questions that determine what comes next. The plan is a conversation.
Where this comes from
Pathstate is built on the adoption model Disruptors Co uses with regulated and risk-sensitive organisations — financial services, government, health. It extends the four-phase roadmap most enterprises already have into eight stages, because four phases are too coarse to act on. Being told you're "in the pilot phase" tells you nothing about Monday.
01The eight stages
Foundation
01Intent
Why are we doing this, and what risk will we accept?
02Ground
Is our data, permissions, and policy layer safe to build on?
Pilot
03Discovery
Where does AI actually create value here, and how risky is each use?
04Proof
Does it work on our real work, and can we prove it?
Scale
05Capability
Can our people apply it to their own work without being told how?
06Production
Can we run it reliably, with monitoring and release discipline?
Embed
07Practice
Is this how the work is done now, or still a thing we do on top?
08Compound
Does each deployment make the next one cheaper and faster?
Where would you honestly place your organisation?
Not where you'd like to be. Free, no account, about nine minutes.