Firebreak
See where your AI use sits, and what it would take to move.
A fixed-scope assessment that places your organisation on one grid, built from evidence, and names the three moves that matter most.
Start the conversationWhat Firebreak is
Firebreak places your organisation on one grid: how consistently AI is used, against how far the controls have got past paper. The placement is built from evidence, not from a self-assessment questionnaire.
The position an organisation gives itself and the position the evidence supports are rarely the same, and the gap between the two is usually the most useful thing the assessment finds.
A firebreak is ground you clear on purpose, cut before the season, so you decide in advance what a fire can take. The argument is in the essay.
What you get
The shakedown run
The evidence pass: artefacts, interviews, and actual inspection of your tooling and tenancies. A control is not asked whether it exists. It is shaken down to see whether it holds.
The report
Your position on the grid, the evidence for it, the gap between where you place yourself and where the evidence puts you, and three moves in the order they should happen. Each move states what it would take to do in-house.
The board one-pager
The square, the finding, the three moves. Readable before a meeting.
The grid
Adoption maturity across the bottom, control maturity up the side. Twelve cells, seven named zones. Naming a position, rather than scoring it, keeps it something you can argue with the evidence.
- Shadow
- Unmanaged, individual AI use with nothing governing it. Organisations do not place themselves here, because from the inside it looks empty.
- Exposed
- Real AI use across the organisation, and no control around it.
- Paper
- A policy exists. Nothing yet enforces it.
- Locked down
- Controls that bind, wrapped around use that never became deliberate. Compliant, and losing ground.
- Narrowing
- Controls are starting to bind and use is maturing. The gap is closing.
- Effective
- Deliberate AI use, and controls that bind, log, and have an owner.
- Resilient
- The end point: effective, and you have already decided what happens when someone else changes the terms.
How it runs
- 1
A short scoping conversation
This is also where the price is set and fixed.
- 2
The shakedown run
The evidence pass across your tooling, tenancies, and the people using them.
- 3
The report and one-pager
Delivered and walked through, so the finding lands before it reaches a board pack.
- 4
What happens next is your call
Run the moves in-house, engage us, or take them elsewhere.
Start to finish, a Firebreak review typically runs a few weeks from the scoping conversation to the walkthrough.
What it costs, and what we promise
Scope depends on the size and complexity of your organisation, so the price is set in the scoping conversation, fixed before any work begins, and does not change after.
Firebreak is priced to be worth running on its own, whether or not you ever engage us again.
Where the recommendations are work we are best placed to deliver, we will propose it. Where they are not, we will point you at the best people to do it, and the report tells you what each move takes so you can run it in-house if you would rather.
Who is behind it
Firebreak is run by Alex Hender, who builds and operates AI systems rather than talking about them. He runs the frontier on his own hardware, at his own cost, so a client never learns those lessons on their members' data.
The work behind it comes from regulated environments: renewables, and member-owned strata. The reasoning runs through the sovereignty series.
If your directors need to be equipped to challenge findings like these, that is what Effective Challenge is for.
Start the conversation
It begins with a short scoping conversation. That is where we set the scope and fix the price.
Start the conversation
Tell us a little about your organisation and we'll set up a short scoping conversation. No payment taken here.