Evidence Express · L0

Nine gates. Three domains. One memo.

Nine gates across three domains — value leakage, AI readiness and decision quality. Each domain is read on its own and reported on its own: an open domain is never offset by an evidenced one, so there is no total, no percentage and no grade. It ends in a memo you can print.

Issuance docket

Reference: —
Issued: Not yet issued
Scope: Self-asserted. 9 gates, three domains, no fieldwork.
Basis: Your answers only. Nothing here has been verified by us.

00 of 09 gates answered

01 · Value leakage

0/3 clearedUntested

Why this domain decides things. Until contribution is known by customer and product, every pricing, mix and cost decision downstream of it is taken on a number nobody has verified.

V1 We know true margin by customer and product, not only gross margin.
Contribution after discounts and cost-to-serve, not gross margin.
V2 Discounts, returns and cost-to-serve are reconciled against contribution.
Reconciled to the ledger, not estimated in a model.
V3 We have a 13-Week cash view we actually decide from.
Decided from, not reported after.

Not tested.

Test it: The Margin Leakage Waterfall — Reconstruct contribution by customer and product before the next pricing move, and read where it actually leaks.

02 · AI readiness

0/3 clearedUntested

Why this domain decides things. An AI programme without a workflow and a baseline cannot produce evidence of value. It can only produce evidence of activity.

A1 AI initiatives are tied to a specific workflow and a baseline.
A named workflow, and a number it has to beat.
A2 Every AI output carries its sources and a named human reviewer.
Sources on the output, a reviewer on the record.
A3 We can tell adoption apart from measured value.
Usage is not value, and neither is enthusiasm.

Not tested.

Test it: The Truth Stack — Place your strongest AI value claim on the stack and see which rung it actually sits on.

03 · Decision quality

0/3 clearedUntested

Why this domain decides things. Without an owner, a written hypothesis and a stop rule, a good outcome and a lucky one are indistinguishable afterwards — and so are the bad ones.

D1 Our biggest decisions have a named owner and a written hypothesis.
One owner, one written claim, before the work starts.
D2 We set a scale threshold and stop rules before we pilot.
Set before the pilot, not negotiated after it.
D3 Value claims carry a baseline, attribution and a measurement window.
Agreed with finance before the claim is made.

Not tested.

Test it: The Evidence-to-Decision Cycle — Put one live decision through the cycle and name the step that is missing.

You can issue a memo on a partial reading. It will state, on its face, which gates were not tested.

What happens next

Bring us the decision. We reply within one working day.

The reading is the first rung. Above it sits the Decision Sprint (L1), Executive Mission (L2) and Decision Cockpit (L3). The engine is chosen by the decision — see the seven decision engines.

Nine gates. Three domains. One memo.

01 · Value leakage

02 · AI readiness

03 · Decision quality

Bring us the decision. We reply within one working day.

Evidence for your next high-stakes decision, starting with nine gates.