A decision audit is structured examination of how you are thinking about a choice — not the choice itself. Like a financial audit, it does not tell you what to buy or sell. It examines whether the numbers support what you already believe, and where they do not. Applied to judgment, it examines whether your frame, evidence, and incentives support your preferred outcome — and what they miss.
That miss is usually the reason you came.
Same human sitting with a hard choice. Two instruments people confuse.
Prefers the uncomfortable question over the helpful answer
One decision, now — frame, evidence, incentives, the unasked
Structured summary without a verdict of what to choose
Coaching and therapy are valuable. They are different jobs. A financial auditor does not also try to heal the CEO's childhood. A decision auditor does not either.
The five dimensions
Every DAUDIT audit runs the same dimensions — not as a checklist you feel, but as a structured pass the instrument applies to your specific situation:
1. The real decision
What you typed is rarely what you are deciding. “Should I hire this person?” may be “Can I tolerate another quarter without someone in the seat?” “Should we raise?” may be “Can I face the board without a growth story?” The audit names the real decision before it entertains the stated one.
2. The frame
What have you treated as fixed that is still open? What criteria are you claiming to use that your behavior contradicts? Frame failures cause more bad outcomes than bad options.
3. The evidence
What is proof, what is hope, what is social pressure, what is fear dressed as prudence? Confidence is separated from evidence — not to embarrass you, to locate where you are guessing.
4. The incentives
Including yours. Especially yours. Where does agreement feel like clarity? Where does speed feel like strategy? Where does delay feel like care when it is avoidance?
5. The unasked questions
The questions whose answers would change the decision — or change which decision you should be making. Often one question dominates. The audit finds it.
Matthews’s separation still applies here: process quality is not outcome quality. A careful close can precede a painful result. A careless bet can get lucky. The audit protects the first; the Ledger scores the second later.
The close
The audit ends in a close — a structured summary, not a recommendation:
- Real decision — restated precisely
- Core tension — the central conflict in your reasoning, in one line
- Blind spots — what you were not weighing
- Decision risk — what happens if you decide without resolving the tension
- Next thinking action — one concrete step that is examination, not execution
There is no “therefore, choose A.” That omission is intentional. The moment the instrument picks a side, you stop examining and start defending. The close is designed to keep the examination alive through the decision — not replace your judgment.
Matthews would separate one further layer many people skip at the end: process quality is not outcome quality. The close protects process examination. Outcome scoring lives later, in the Ledger — after contact with the world.
Why independence matters
You cannot audit your own books with the same rigor an independent auditor can. You know where the bodies are buried. You also know where you want bodies to be buried — in the decisions you have already leaned toward.
DAUDIT is built to default toward the uncomfortable question, not the helpful answer. That is the opposite of your AI's agreeableness — and the opposite of most human conversations, where your friends want you to be happy and your colleagues want you to move.
Who it is for
Founders, investors, operators — and anyone facing a life decision with capital-grade stakes: care, relocation, partnership, exit from a relationship or role. The two-audience principle is not marketing. The method is the same because the failure modes are the same: closed frames, flat confidence, and incentives nobody named.
Try it on something real
Generic examples teach little. Open an audit on a decision you are actually facing — not a hypothetical. The examination only works with skin in the game.
For how decisions compound across time in the instrument, read Why DAUDIT remembers. For the AI contrast, start with [Your AI agrees with you](/blog/your-ai-agrees-with-you).