A citizen can appeal the decision. You have to be able to reconstruct it.
Administrative decisions carry a right of review. That right is only real if the decision can be reproduced exactly as it was made — including which rule version applied on the day. This scenario shows that constraint enforced.
What you are already obliged to produce
Public-sector obligations attach to the decision, not to the tool that helped make it.
| Authority | What it obliges you to produce |
|---|---|
| Privacy Act §552a | Accounting of disclosures, and a record's accuracy and relevance maintained for any determination made about an individual. |
| NIST SP 800-53 AU | Audit event capture, content, and protection sufficient to reconstruct what occurred and under which configuration. |
| FISMA | Agency-wide information security programme with continuous monitoring and documented control effectiveness. |
| FOIA | Records subject to disclosure on request — which presumes they exist, are locatable, and are intelligible to a reviewer. |
Why the usual answer does not close
A determination nobody can reproduce is a determination nobody can appeal.
Rule versions change. A decision reviewed a year later must be evaluated against the rules in force that day, not the rules in force at review.
Audit records assembled from application logs record what the system reported, which is not the same as what the system decided.
Procurement requires demonstrated control effectiveness before deployment, and a probabilistic guardrail cannot demonstrate the same behaviour twice.
One action at the boundary
Every disposition is bound to the policy hash in force at decision time, so a review reconstructs the decision rather than approximating it.
Candidate action
benefit.determination.finalise · appeal window still open
Finalising while the appeal window is open would foreclose a right the applicant still holds. The block is not a judgement about the merits — it is an ordering constraint the policy engine cannot bypass, recorded with the rule version that produced it so a reviewer sees the same state the decision saw.
The same measurements, whatever the sector
These are figures from the internal technical evidence report, not projections modelled for this scenario. They describe one pipeline, so they do not change when the mandate does.
0 / 18
Attack episodes passed
Stage 0 SHADOW evaluation of action traces authored by real generative planners. None of nine held-out benign episodes was blocked.
0 / 10
Prohibited cases passed, dual reference
Controlled ablation on the same held-out set. A single-reference gate passed four of ten under identical calibration.
0.47 ms
Mean decision latency
Across 5,000 measured decisions; 1.15 ms at p99. The gate itself completes in 258 nanoseconds.
1,000 / 1,000
Identical digests on replay
Same input, pinned environment. Ten thousand ledger records re-verified in 0.195 seconds.
What this scenario is not
Read this before you quote it
- An illustrative scenario, not a customer engagement. No client is named because none is being described.
- No deployment in this sector is claimed, and no regulator has reviewed or endorsed this material.
- The figures are prototype measurements on one commodity workstation, not production or distributed results.
- Nothing here has been independently reproduced by a third party.
- The authorities cited describe the obligation you carry — not a determination that we satisfy them.
- Operational accuracy on a validated sector corpus remains outstanding; a pilot requires your own reference data.
Check the numbers before you trust the scenario
Every figure above is drawn from the technical evidence page, where the same measurements appear with their tail distribution and their stated limits.