AI Decision Log Audit — We Considered A, Chose B, Because C
AI decision log audit: Edmund Ng's journey spoke on governed AI, harness testing, and Vibe Coding for solo founders. Explore.
Published Updated 13 min read
auditable-aidecision-loggovernance
AI decision log audit matters when you move from demo velocity to production scrutiny. This article is Edmund Ng's field notes on AI tradeoff record, harness discipline, and the journey toward auditable AI—written for solo founders and system rule designers who cannot afford silent regressions.
Continue with these journey spokes.
Continue with these journey spokes.
Continue with these journey spokes.
Continue with these journey spokes.
Continue with these journey spokes.
Continue with these journey spokes.
Building Auditable AI Systems · What Is an Evidence Chain in AI · Knowledge Graph for AI Compliance
On this page
- What — AI decision log audit — AI tradeoff record — Decision Log defined
- Why — decision log audit trail — outcomes without reasons fail audit
- When — AI tradeoff record — log a decision
- Where — decision log audit trail — Decision Log in the artifact family
- How — AI tradeoff record — write one entry
- 是什么 — extended AI tradeoff record — decision log audit trail
- 为什么 — extended decision log audit trail — AI tradeoff record
Key takeaways
- AI decision log audit needs written rules—not hero prompts alone.
- AI tradeoff record keeps demo speed from becoming production regret.
- Harness discipline connects this spoke to the wider governed production journey.
- Cross-link Phase docs, Harness retests, and written tradeoff logs before calling work done.
Takeaways above anchor the rest of this spoke.
What — AI decision log audit — AI tradeoff record — Decision Log defined
A Decision Log is a human-readable record of options considered, option chosen, and documented reason at the moment an architectural or product decision is made.
Template:
We considered A, chose B, because C.
| Field | Meaning |
|---|---|
| A | Credible alternative(s) — including "do nothing" or "defer" |
| B | The committed choice |
| C | Evidence-linked reason — constraints, risks, measurements, policy |
This is not a meeting minute. It is a constitutional artifact — same category as Phase documents for session memory, but focused on tradeoffs, not task checklists.
Structured exports and harness retests matter more than demo velocity when reviewers ask for evidence.
Structured exports and harness retests matter more than demo velocity when reviewers ask for evidence.
Governed exports and harness checkpoints prevent demo velocity from collapsing under review.
In the What layer of this Act 3 auditable AI spoke, teams work from an operational contract—not a marketing label. Governed exports and harness checkpoints prevent demo velocity from collapsing under multi-axis review or compliance questions. A practical test for what belongs in an AI decision audit: what is frozen before agents sweep, what gets logged at tradeoff time, and which Harness retest proves behavior instead of UI luck. Edmund Ng's field notes emphasize exportable rules and written logs so six-month-later auditors can follow the chain—that is the same fast AND governed bridge Acts 1–3 teach.
Why — decision log audit trail — outcomes without reasons fail audit
Teams often document what shipped and reconstruct why months later. Under scrutiny, reconstructed stories collapse — especially when models or staff rotate.
Decision Log prevents:
| Failure | How Decision Log helps |
|---|---|
| Post-hoc rationalization | Reason recorded at decision time |
| Silent regression | Rejected option A stays rejected with documented C |
| Model amnesia | Future LLMs read why architecture looks this way |
| Audit theater | Tradeoffs visible without a forensic chat archaeology project |
- Evidence chain → path visibility
- Decision Log → ownership and tradeoff visibility
- Harness + gates → proof the system behaves as documented
A Decision Log records what we considered, what we chose, and why. It is not written after the fact.
Structured exports and harness retests matter more than demo velocity when reviewers ask for evidence.
Governed exports and harness checkpoints prevent demo velocity from collapsing under review.
In the Why layer of this Act 3 auditable AI spoke, teams work from an operational contract—not a marketing label. Governed exports and harness checkpoints prevent demo velocity from collapsing under multi-axis review or compliance questions. A practical test for why real time AI tradeoffs need logs: what is frozen before agents sweep, what gets logged at tradeoff time, and which Harness retest proves behavior instead of UI luck. Edmund Ng's field notes emphasize exportable rules and written logs so six-month-later auditors can follow the chain—that is the same fast AND governed bridge Acts 1–3 teach.
When — AI tradeoff record — log a decision
Always log when:
- Choosing between architecture patterns (sync vs async orchestration, gate strictness, tenant isolation model)
- Rejecting a feature path after multi-axis review findings
- Changing Constitution, Framework, or Instruction Governance rules
- Closing a 10/80/10 POST cycle with rule writeback
Skip when:
- Typo fixes with zero tradeoff surface
- Experiments explicitly labeled discard — but log the discard decision if scope was serious enough to confuse future readers
Status honesty: Decision Log is a formal layer (May 2026) in Edmund's stack. Propagation to all historical Phase documents is in progress — the method is stable; backfill is ongoing.
Structured exports and harness retests matter more than demo velocity when reviewers ask for evidence.
Structured exports and harness retests matter more than demo velocity when reviewers ask for evidence.
Governed exports and harness checkpoints prevent demo velocity from collapsing under review.
In the When layer of this Act 3 auditable AI spoke, teams work from an operational contract—not a marketing label. Governed exports and harness checkpoints prevent demo velocity from collapsing under multi-axis review or compliance questions. A practical test for how to write an AI decision log: what is frozen before agents sweep, what gets logged at tradeoff time, and which Harness retest proves behavior instead of UI luck. Edmund Ng's field notes emphasize exportable rules and written logs so six-month-later auditors can follow the chain—that is the same fast AND governed bridge Acts 1–3 teach.
Where — decision log audit trail — Decision Log in the artifact family
| Artifact | Granularity | Purpose |
|---|---|---|
| Decision Log | One tradeoff sentence | Human-readable why |
| Phase Gap Ledger | Per-phase findings | Build readiness |
| §22 appendices / dossiers | Wave checkpoints | Append-only program memory |
| Evidence chain | Runtime custody | Source-to-conclusion proof |
Relationship: Same spirit, different granularity. A Decision Log entry might reference an Evidence Snapshot id or a harness run — but stays readable without opening code.
Connects to Stage A/B: Stage B mutations should cite the Decision Log (or gate envelope derived from it) — read-before-write for commits.
Structured exports and harness retests matter more than demo velocity when reviewers ask for evidence.
Structured exports and harness retests matter more than demo velocity when reviewers ask for evidence.
Governed exports and harness checkpoints prevent demo velocity from collapsing under review.
In the Where layer of this Act 3 auditable AI spoke, teams work from an operational contract—not a marketing label. Governed exports and harness checkpoints prevent demo velocity from collapsing under multi-axis review or compliance questions. A practical test for what belongs in an AI decision audit: what is frozen before agents sweep, what gets logged at tradeoff time, and which Harness retest proves behavior instead of UI luck. Edmund Ng's field notes emphasize exportable rules and written logs so six-month-later auditors can follow the chain—that is the same fast AND governed bridge Acts 1–3 teach.
How — AI tradeoff record — write one entry
- Name the decision — one line scope (e.g. "Tenant isolation for cross-domain handoff")
- List A — at least one real alternative, including deferral
- State B — the commit
- Write C — constraints, metrics, policy refs (abstract in public writing)
- Link evidence — snapshot id, review marker doc, or Phase doc section
- Append only — supersede with new entry if B changes; never edit history silently
Example (abstract, Pattern layer):
We considered parallel unbounded agent fan-out, chose governed sequential orchestration with read-only parallelization under policy, because audit replay and tenant isolation require a single primary execution path per turn.
Rule writeback: when multi-axis review finds a blindspot, the heuristic becomes permanent — Decision Log captures that policy change too.
Structured exports and harness retests matter more than demo velocity when reviewers ask for evidence.
Governed exports and harness checkpoints prevent demo velocity from collapsing under review.
In the How layer of this Act 3 auditable AI spoke, teams work from an operational contract—not a marketing label. Governed exports and harness checkpoints prevent demo velocity from collapsing under multi-axis review or compliance questions. A practical test for why real time AI tradeoffs need logs: what is frozen before agents sweep, what gets logged at tradeoff time, and which Harness retest proves behavior instead of UI luck. Edmund Ng's field notes emphasize exportable rules and written logs so six-month-later auditors can follow the chain—that is the same fast AND governed bridge Acts 1–3 teach.
是什么 — extended AI tradeoff record — decision log audit trail
Governed builders treat written rules, frozen snapshots, and harness retests as production requirements—not optional polish after a green demo. Edmund Ng's journey from non-programmer Vibe Coding to auditable AI systems shows why structure beats model churn when stakeholders ask how you decided, what you rejected, and what evidence you can export tomorrow.
Governed builders treat written rules, frozen snapshots, and harness retests as production requirements—not optional polish after a green demo. Edmund Ng's journey from non-programmer Vibe Coding to auditable AI systems shows why structure beats model churn when stakeholders ask how you decided, what you rejected, and what evidence you can export tomorrow.
Structured exports and harness retests matter more than demo velocity when reviewers ask for evidence.
Governed exports and harness checkpoints prevent demo velocity from collapsing under review.
In the 是什么 layer of this Act 3 auditable AI spoke, teams work from an operational contract—not a marketing label. Governed exports and harness checkpoints prevent demo velocity from collapsing under multi-axis review or compliance questions. A practical test for how to write an AI decision log: what is frozen before agents sweep, what gets logged at tradeoff time, and which Harness retest proves behavior instead of UI luck. Edmund Ng's field notes emphasize exportable rules and written logs so six-month-later auditors can follow the chain—that is the same fast AND governed bridge Acts 1–3 teach.
为什么 — extended decision log audit trail — AI tradeoff record
Governed builders treat written rules, frozen snapshots, and harness retests as production requirements—not optional polish after a green demo. Edmund Ng's journey from non-programmer Vibe Coding to auditable AI systems shows why structure beats model churn when stakeholders ask how you decided, what you rejected, and what evidence you can export tomorrow.
Governed builders treat written rules, frozen snapshots, and harness retests as production requirements—not optional polish after a green demo. Edmund Ng's journey from non-programmer Vibe Coding to auditable AI systems shows why structure beats model churn when stakeholders ask how you decided, what you rejected, and what evidence you can export tomorrow.
Structured exports and harness retests matter more than demo velocity when reviewers ask for evidence.
Governed exports and harness checkpoints prevent demo velocity from collapsing under review.
In the 为什么 layer of this Act 3 auditable AI spoke, teams work from an operational contract—not a marketing label. Governed exports and harness checkpoints prevent demo velocity from collapsing under multi-axis review or compliance questions. A practical test for what belongs in an AI decision audit: what is frozen before agents sweep, what gets logged at tradeoff time, and which Harness retest proves behavior instead of UI luck. Edmund Ng's field notes emphasize exportable rules and written logs so six-month-later auditors can follow the chain—that is the same fast AND governed bridge Acts 1–3 teach.
Summary
AI decision log audit on Edmund Ng's journey means shipping with AI tradeoff record, harness retests, and evidence-friendly decisions—not one-off prompts. Models change; written rules, exportable snapshots, and governance patterns endure.
Governed builders treat written rules, frozen snapshots, and harness retests as production requirements—not optional polish after a green demo. The journey from non-programmer Vibe Coding to auditable AI shows why structure beats model churn when stakeholders ask how you decided, what you rejected, and what evidence you can export tomorrow.
How to write an AI decision log
Edmund Ng treats each long-tail question as a production gate: freeze the spec, log the tradeoff, and prove behavior with Harness retests—not demo clicks alone.
Governed builders treat written rules, frozen snapshots, and harness retests as production requirements—not optional polish after a green demo. The journey from non-programmer Vibe Coding to auditable AI shows why structure beats model churn when stakeholders ask how you decided, what you rejected, and what evidence you can export tomorrow.
What belongs in an AI decision audit
Edmund Ng treats each long-tail question as a production gate: freeze the spec, log the tradeoff, and prove behavior with Harness retests—not demo clicks alone.
Solo founders in Malaysia and APAC often face professional scrutiny early. Externalizing Phase documents, Decision Logs, and smoke tiers before the demo invitation arrives is cheaper than rebuilding trust after a silent regression reaches a customer walkthrough.
Why real time AI tradeoffs need logs
Edmund Ng treats each long-tail question as a production gate: freeze the spec, log the tradeoff, and prove behavior with Harness retests—not demo clicks alone.
Role separation matters: builder models may sweep diffs, but frontier models should audit frozen snapshots. Mixing those hats in one chat thread is how teams lose reproducibility and inherit context debt that no IDE upgrade fixes.
FAQ
What is AI decision log audit?
Edmund Ng answers with structure first: freeze specs, separate builder and frontier roles, and prove behavior with Harness—not demo clicks. Written rules, Phase documents, and Decision Logs let teams explain tradeoffs months later without reconstructing chat history.
Governed builders treat written rules, frozen snapshots, and harness retests as production requirements—not optional polish after a green demo. The journey from non-programmer Vibe Coding to auditable AI shows why structure beats model churn when stakeholders ask how you decided, what you rejected, and what evidence you can export tomorrow.
How to write an AI decision log?
Edmund Ng answers with structure first: freeze specs, separate builder and frontier roles, and prove behavior with Harness—not demo clicks. Written rules, Phase documents, and Decision Logs let teams explain tradeoffs months later without reconstructing chat history.
Solo founders in Malaysia and APAC often face professional scrutiny early. Externalizing Phase documents, Decision Logs, and smoke tiers before the demo invitation arrives is cheaper than rebuilding trust after a silent regression reaches a customer walkthrough.
What are the required items in an AI decision audit?
Edmund Ng answers with structure first: freeze specs, separate builder and frontier roles, and prove behavior with Harness—not demo clicks. Written rules, Phase documents, and Decision Logs let teams explain tradeoffs months later without reconstructing chat history.
Role separation matters: builder models may sweep diffs, but frontier models should audit frozen snapshots. Mixing those hats in one chat thread is how teams lose reproducibility and inherit context debt that no IDE upgrade fixes.
Why does real time AI tradeoffs need logs?
Edmund Ng answers with structure first: freeze specs, separate builder and frontier roles, and prove behavior with Harness—not demo clicks. Written rules, Phase documents, and Decision Logs let teams explain tradeoffs months later without reconstructing chat history.
Governed builders treat written rules, frozen snapshots, and harness retests as production requirements—not optional polish after a green demo. The journey from non-programmer Vibe Coding to auditable AI shows why structure beats model churn when stakeholders ask how you decided, what you rejected, and what evidence you can export tomorrow.
Why does AI tradeoff record matter for solo founders?
Edmund Ng answers with structure first: freeze specs, separate builder and frontier roles, and prove behavior with Harness—not demo clicks. Written rules, Phase documents, and Decision Logs let teams explain tradeoffs months later without reconstructing chat history.
Solo founders in Malaysia and APAC often face professional scrutiny early. Externalizing Phase documents, Decision Logs, and smoke tiers before the demo invitation arrives is cheaper than rebuilding trust after a silent regression reaches a customer walkthrough.
When should teams freeze specs before agent sweeps?
Edmund Ng answers with structure first: freeze specs, separate builder and frontier roles, and prove behavior with Harness—not demo clicks. Written rules, Phase documents, and Decision Logs let teams explain tradeoffs months later without reconstructing chat history.
Role separation matters: builder models may sweep diffs, but frontier models should audit frozen snapshots. Mixing those hats in one chat thread is how teams lose reproducibility and inherit context debt that no IDE upgrade fixes.
About the author
Edmund Ng — Malaysia-based solo founder, AI systems architect, and system rule designer. He ships governed AI with Vibe Coding, harness engineering, and auditable evidence chains. About · Projects · LinkedIn.