DECISION SYSTEM

Risk and Decision

Keep risks live as evidence changes, then preserve the reasons behind your own decision.

0 open 0 triggered 0 due

ALWAYS-ON CROSS-MODULE REVIEW

AI Executive Briefing

0 source records · 0 accepted agent drafts. Briefings remain review-only.

Step 1 Briefing prepared Assembled across evidence, compound, valuation and decision from records you already hold. Reading is all it does.
Step 2 Draft in the Review Lane Lands in the agent drawer, dated and editable. It is not in your assessment, and it is not on this page.
Step 3 · only if you accept Your assessment record Reached by you accepting the draft, in your wording. Reject it and nothing is written, anywhere, ever.

There is no route that skips step 2. The briefing has no write path into the assessment: it cannot create a risk, move a probability, set a status, pin a record or fill a field. Everything it produces stops in the Review Lane and waits. That is what review-only means, and it is true on the day it has nothing to say as well as on the day it has four drafts ready.

State shown

Cross-module review

No deterministic cross-module inconsistency rule fired. Continue reviewing unresolved evidence and contrary cases.

Evidence briefingReads your evidence board and the provenance tag on every line.

Not drafted

Compound briefingReads your owner-earnings, ROIC and runway build-ups, and your separate confidence on each.

Not drafted

Valuation briefingReads your intrinsic-value assumptions and the ranges you stated for each.

Not drafted

Decision briefingReads your thesis fields, your risk register and your invalidators.

Not drafted

DESIGNED The live capture held no source records and fired no rule, so it could only show the panel with nothing in it. Below is the same panel with material behind it and one rule fired — written, not quoted.

Cross-module review 1 rule fired

CM-02 — an assumption used in one module has no supporting entry in another. Your Valuation extension carries a ten-year revenue growth assumption. Your Compound Analysis growth-runway entry names expansion vectors covering three years and records the runway beyond that as not established. Both entries are yours, both are still saved exactly as you left them, and this review changed neither.

The rule reports the difference and stops. It does not say which of your two entries is the wrong one, it does not reconcile them, and it has no opinion about FLWS. Continue reviewing unresolved evidence and contrary cases.

Evidence briefingReads your evidence board and the provenance tag on every line.

Draft waiting · 24 Jul

Compound briefingReads your owner-earnings, ROIC and runway build-ups, and your separate confidence on each.

Accepted 22 Jul · your edit

Valuation briefingReads your intrinsic-value assumptions and the ranges you stated for each.

Rejected 21 Jul · nothing written

Decision briefingReads your thesis fields, your risk register and your invalidators.

Not drafted

Four rows, four dispositions, and only one of them is a record. Not drafted means the briefing was never asked for. Draft waiting means it exists in the agent drawer and nowhere else. Rejected means it was read and thrown away, and the word for what it left behind is nothing. Accepted is the only one that reached the assessment, it reached it as your edited wording, and it is counted in the “accepted agent drafts” figure at the top of this panel so the count is always the number of times you said yes.

DYNAMIC REGISTER

Risks you are monitoring

Probability, impact, and status change only when you save a new version.

CAPTURED This is the register as the running product shows it today: nothing recorded, and no seeded industry risks standing in for the ones you have not written yet.

No user-authored risks yet

Add a risk or ask the Skeptic for reviewable questions. Agent output never writes here directly.

Why the register starts blankEmpty is a true state

A borrowed risk is not one you will recognise when it arrives. The Skeptic can hand you questions to consider, but a question is not a register line — you write the line, in your wording, or the register stays empty.

Agent output never writes here directly. Whatever an agent returns arrives as a proposal you accept, edit or reject. Accepting is what creates the row, and your wording is what gets stored.

DESIGNED The live capture showed only the empty register, so every row, rating and status below is a design proposal — not observed product content. The counter strip above moves with it.

Risk register4 recordedOne material risk per line

Risk · your wording	Probability	Impact	Status	Saved version
Gifting demand normalises faster than the cost base can followThe same argument as my bear case, written as a line I can watch.	3 / 5	4 / 5	Open	v3 · 22 Jul
Customer acquisition gets more expensive faster than repeat ordering offsets itThe one thing I said I cannot yet judge from the outside.	2 / 5	4 / 5	Open	v1 · 12 Jul
Fulfilment and delivery cost inflation the price list cannot absorb in a peak weekPeak weeks are when the margin is made and when the cost is least controllable.	3 / 5	2 / 5	Triggered	v2 · 24 Jul
Portfolio breadth turns into overhead rather than reachI have not decided yet whether this is material.	2 / 5	3 / 5	Open	v1 · 12 Jul

Probability, impact, and status change only when you save a new version. Typing in a field does not move the register; saving does, and the previous version stays readable. The two ratings are never multiplied into a severity score, the rows are never re-ordered by anything the app worked out, and the register is shown in the order you wrote it.

DYNAMIC REGISTER

Triggers on the risks you wrote

Each trigger belongs to one risk. Firing is a fact you record; what it means is still yours to decide.

DESIGNED Not captured. The running product showed an empty Risk register and never rendered this tab with content, so the dashboard below is a design proposal built from the designed register rows.

Trigger dashboard4 triggers · 1 firedDerived from your register

Trigger · your wording	From	Where you watch it	Last checked	State
Two consecutive quarters where gross margin falls while order volume holdsSeparates a cost problem from a demand problem.	Risk 1	Segment revenue and gross margin in the 10-Q	24 Jul	Not fired
Marketing spend rises for two periods running without a matching move in ordersVisible in the filings long before it shows up in the margin.	Risk 2	Marketing expense against order and customer counts	24 Jul	Not fired
Fulfilment and delivery cost per order rises across two peak seasonsRecorded as fired by you on 24 Jul, against the holiday-quarter cost commentary.	Risk 3	Cost of revenue commentary and the holiday-quarter call	24 Jul	Fired · 24 Jul
A brand in the portfolio is impaired, wound down or divestedThe disclosure usually lands before the strategy changes in public.	Risk 4	8-K filings and impairment disclosures in the 10-K	24 Jul	Not fired

What a fired trigger is, and is not. It is the observable event you named, recorded as having happened, on the date you recorded it. It is not a sell signal, a downgrade or a change of status: risk 3 stays exactly as you last saved it until you save a new version. The dashboard reports; you decide.

Reviews due1 dueDates you set

Risk 3 — fulfilment cost inflationreview due 31 Jul

Risk 1 — gifting demand normalises22 Oct

Risk 2 — acquisition cost12 Jan

Risk 4 — portfolio breadth12 Jan

A review date is a reminder you set, not a deadline the app imposes. Nothing expires, nothing is auto-closed, and a missed review is shown as missed rather than quietly cleared.

CoverageCompleteness only

✓Every risk has a trigger4 of 4

✓Every trigger names where to look4 of 4

✓Every trigger has a check date4 of 4

Fired trigger reviewedoutstanding

These check whether your work exists, not whether it is right. A risk without a trigger is a mood; a trigger without a place to look is a wish. Neither check is an opinion about the company.