The Backbrief: the most honest document in the room

Most strategy dies in the gap between "I told them" and "they understood."

A leader sets a direction — clearly, they think. It cascades down. Everyone nods. Months later the result comes back subtly wrong, and the leader is baffled: I was explicit. They were. Explicit instructions just aren't the same thing as shared intent, and the difference only shows up after it's too late to fix cheaply.

The military worked this out a long time ago. It's called mission command, or Auftragstaktik: you give people the intent — what must be achieved and why — and let them work out the how, because the people closest to the problem will always know the terrain better than you do. But mission command has a safety mechanism that most business adaptations quietly drop. It's called the backbrief, and it's the single most useful artifact I've built into my strategy tooling.

What a backbrief actually is

A backbrief is the subordinate restating the challenge in their own words, back up the chain, before they start. Not "yes, understood." A real one declares:

The effects they're actually going to pursue.
Hard constraints vs. soft ones — the lines that cannot move, separated from the preferences that can.
Anti-goals — the ways they refuse to win.
An explicit autonomy boundary — "here is what I will decide on my own, and here is what I will escalate, and when."
The risks, dependencies and blockers they can already see.

When you read a good one, two things happen. You find out whether your intent survived the journey. And you find out where the person disagrees with you — which, if you've built the right culture, is the most valuable thing in the document. Disagreement is signal. A backbrief is where it gets to surface before it becomes a failure.

What it looks like when it's real

Here's the shape of a good one, from a worked example in the system — a shipping line told to decarbonise a trade lane.

The weak version reads: "I will cut emissions. I will not fail. I'll decide everything and escalate nothing." That's not alignment, it's noise dressed as confidence.

The real version is specific enough to be uncomfortable:

Hard constraint: schedule reliability stays at or above 97.5%. Non-negotiable. Customers leave below that, and a decarbonised lane with no cargo is not a win.
Anti-goal: I will not hit the fuel target by quietly cutting service frequency. (The obvious cheat, named and ruled out up front.)
Autonomy boundary: I'll decide on vessel speed profiles and routing myself. I'll escalate anything that touches customer contracts.
Early-warning trigger: if reliability drops to 98.5%, I raise it — before the 97.5% hard line breaks, not after.

That last point is the tell of a mature backbrief. The escalation fires while there's still room to act, not once the gate has already failed. It turns "I'll let you know if something goes wrong" into a specific, measurable promise.

Why I made the AI do this, and not more

It would be easy to build a tool that writes the backbrief for you. I deliberately didn't. The AI facilitates — it asks the questions, checks the document is actually specific, flags the anti-goals that are too vague to constrain anything. But it never decides the strategic direction, and it never puts words in the subordinate's mouth. The authority, and the author's voice, stay human.

Because the entire point of a backbrief is that the person who has to deliver it owns it. An AI-generated backbrief the person didn't actually think through is worse than none at all — it's the illusion of alignment, which is exactly the thing that kills strategies in the first place.

The takeaway

If your directions keep coming back subtly wrong, the problem usually isn't that you were unclear. It's that you never heard it said back. Ask for a backbrief — effects, hard constraints, anti-goals, where they'll act alone and where they'll escalate — and read it for where they disagree with you.

That document is the cheapest insurance you will ever buy against a strategy that nods its way to failure.

I'm Tete Mensa-Annan. I spent years turning groups into self-sustaining teams — at Microsoft, then running an engineering platform org at Novo Nordisk — on one principle: intent over instruction, autonomy over control. If your strategy keeps dissolving on its way down the organisation, I'm happy to talk about where the dialogue is missing.