Four stages, in that order. Most owners start at Automate — which is how you end up automating work that should not exist.
The point at which your ability to create, execute and scale new value consistently exceeds the rate at which disruption erodes the value you already have.
Try raising Execution. When your competitors have the same tools, execution speed cancels out of the inequality — it raises the rate of disruption at exactly the rate it raises your output. Only Innovation and Adoption move where you sit.
Execution got cheap. Judgment did not.
Where the business actually decides, what it can demonstrate rather than believe, and where you are the constraint.
Is it strategically sound — the right move, for the right reasons, aligned to where the business is actually going?
↳ Innovation
Is it feasible and ready — the data, the capability and the operating reality to execute without breaking?
↳ Execution
Will it be trusted and adopted, or drift the moment it leaves the room? A decision people won’t run isn’t validated.
↳ Adoption
Pass two of three and you do not have a validated decision — you have one with a known failure mode nobody wrote down.
Pick a stage above.
Put AI where it belongs — and remove the work that shouldn’t exist before anything gets built on top of it.
Delete before you automate.
Automating work that shouldn’t exist is the most expensive way to keep it — you pay for the tool, the old process, and the people maintaining both. Nobody else will tell you this, because deleting work is revenue they lose.
Five steps that turn a business outcome into validated execution. It begins with where you’re going, not the technology.
Start with the business outcome and the decision that gets you there, not the tool.
Surface the real decision, the current state and the hidden assumptions underneath it.
Validate the decision — criteria, ownership and risk — using the VALID™ framework.
Design the strategy, AI and systems around the validated decision. Decide what to hand AI.
Build, adopt and govern, so the decision ships and compounds in Validated OS.
Chris takes the five steps in order — where the decision gets validated, and where AI enters.
A validated decision nobody adopts is a document. This is the stage that decides whether anything actually changes.
Your managers stop interpreting and start repeating. Same answer in every room, without you in it.
While the manual route stays available, people use it. Closing it is the single most skipped step in every rollout we see.
Not licences issued. Whether the decision is being made the new way, by the people who own it.
Last, and only what has earned it. Four settings, chosen per decision — “out of the loop” is one option of four, not the goal.
Judgment stays with the team. Right wherever context, relationships or accountability carry the weight.
The system gathers and sets out the options. A person makes the call every time.
The system handles the routine. Your team watches the exceptions and steps in.
End to end, inside limits you set in advance and can change whenever you choose.
Then claim the return.
Decide what the freed capacity is for, retire what it replaced, and measure against the baseline you took in Assess. A gain nobody claims stays on a slide.
90 days · credited in full if you continue
Assess, and the first half of Apply & Delete. Gets the decisions off your desk.
Typically nine months · all four
End to end. Anything invested with us in the previous twelve months comes off.