| name | kaizen-engine-and-product-improvement |
| description | Use when auditing or improving the finance doctrine engine or any accounting, control, reporting, tax, budget, reconciliation, or finance-system product it produces. |
| metadata | {"portable":true,"compatible_with":["claude-code","codex"]} |
Kaizen Engine and Product Improvement
Acknowledgement: Shared by Peter Bamuhigire, techguypeter.com.
Last reviewed: 2026-08-04
Next review due: 2026-11-04
Use When
- Auditing this doctrine engine or a finance/accounting artefact, control, report, model, integration, or system.
- Turning close, reconciliation, control, audit, incident, user, or reviewer findings into a tested standard improvement.
Do Not Use When
- A single skill safety or statutory-source check is sufficient.
- Current IFRS, tax, statutory, or regulatory claims have not been routed through current source registers and verification.
Prerequisites
Read the accounting-finance doctrine, finance release gate, applicable source register, and the Digital Research Skills Engine source-evaluation and source-verification skills.
Inputs
| Artefact | Source/provider | Required? | Purpose | If absent |
|---|
| Scope/reporting basis, source register, transaction/control evidence, product type, current score, risks, reviewer, and target | Finance project and doctrine | yes | Set audit scope and improvement target | Stop or mark unassessed |
Workflow
- Read the local adoption plan, doctrine, quality gate, and portfolio standard.
- Inventory doctrine, skills, source registers, control matrices, examples, invariants, tests, and cross-engine routes.
- Score each applicable dimension and product type. Publish
min(raw score, 65); never use the cap to waive a control failure.
- Audit accounting correctness, completeness, cut-off, reconciliation, controls, segregation of duties, privacy/security, explainability, usability, evidence, and handoff.
- Write a 95/100 remediation plan with owner, root cause, experiment, acceptance evidence, residual risk, and rollback.
- Run a small control/process improvement. If evidence or controls fail, stop, recover the safe state, and re-perform.
- Standardise only after evidence; update doctrine/reference/fixture/test, re-test, and schedule re-audit.
Outputs
| Artefact | Consumer | Acceptance condition |
|---|
| Capped audit scorecard, control/evidence gap register, 95/100 plan, experiment result, and residual-risk decision | Controller, reviewer, and release owner | Evidence, owner, control impact, acceptance proof, rollback, and next review are explicit |
Evidence Produced
| Evidence | Format | Acceptance condition |
|---|
| Source checks, invariants, re-performance, report rebuild, independent review, and residual-risk record | Validator output, workpaper, or evidence pack | Another reviewer can reproduce the conclusion and verify the control remained effective |
Capability Contract
Read and search are required. Audits are read-only by default; posting, control, source, or production changes require explicit authority and permission. Route current facts to Digital Research.
Degraded Mode
If the source register, transaction population, control evidence, reviewer, or validator is unavailable, return a qualified result, mark the gap not assessed, and do not certify compliance or release readiness.
Decision Rules
| Condition | Action | Failure or risk avoided |
|---|
| A control objective would be removed to save time | Reject and redesign for waste reduction | Misstatement or fraud |
| A posting or report change lacks re-performance | Stop, test invariants, and rebuild | Silent accounting drift |
| A change passes controls and independent review | Standardise it and schedule re-audit | Repeated exceptions |
Acceptance Evidence
Run the doctrine validator, source-register checks, accounting invariants, finance release gate, changed-skill validation, and independent controller review. Record residual risk and the next review date.
Quality Standards
Finance doctrine remains authoritative. Kaizen cannot weaken period locks, audit trails, segregation of duties, reconciliation, statutory verification, privacy, or approval controls. A faster process that increases misstatement or fraud risk is a failed experiment.
Mandatory 65-to-95 gate
The first pass is an initial analysis, not a certification: calculate the raw result,
publish min(raw_score, 65), and keep control failures and unassessed evidence
separate from the score. Only then may a reversible improvement cycle target 95/100.
Each proposed change must state its root cause, control guardrail, owner, measure,
stop/rollback rule, re-performance evidence, residual risk, and re-audit date.
Run the gate at engine level (doctrine, skills, controls, references, validators,
and handoffs) and product level (report, reconciliation, budget, control, or finance
system). A product cannot inherit a passing result from the engine.
Anti-Patterns
- Treating lean as fewer controls. Fix: remove waste, not control objectives.
- Changing a posting rule without re-performance. Fix: run invariants and report rebuilds.
- Using historical book claims as current statutory facts. Fix: use dated source registers.
- Closing remediation without residual-risk owner. Fix: keep open or formally accept.
- Hiding exceptions in a score. Fix: record blockers separately.
Worked Example
If a reconciliation workaround removes a repeated manual step but weakens the audit trail, reject the change, retain the control, and test an automated evidence-preserving alternative.
References
- Local adoption plan
- Portfolio standard:
C:\wamp64\www\digital-research-engine\docs\continuous-improvement\portfolio-kaizen-standard-2026-08.md
skills/06-close-consolidation-and-reporting/finance-module-audit/
governance/finance-accounting-quality-gate.md