Auditability: Why AI and Humans Must Share the Same Record
The two-log problem
Most systems that involve AI create two separate records. There is the operational log: who did what, when, to which resource. And there is the AI log: what the model was asked, what it produced, what confidence score it assigned. These logs live in different systems, use different schemas, and are maintained by different teams.
When something goes wrong, the first task is reconciliation. "The AI recommended this at 14:32. The operator approved it at 14:35. But the execution log shows a different action at 14:37. What happened in between?" This reconciliation consumes hours and often produces inconclusive results because the two records were never designed to align.
What auditability means here
Auditability means AI and human actions are recorded in the same step instances, with the same structure, the same timestamps, and the same artifacts. There is no separate AI log. When the AI proposes an action, that proposal is a record in the operational trail. When a human approves or dismisses it, that decision is the next record in the same trail.
The result is a single, linear sequence of events:
- Step instance created (by dispatcher)
- AI proposal generated (with reasoning and context)
- Human decision recorded (approved, modified, or dismissed)
- Action executed (or not)
- Outcome recorded
Every entry has the same format. Every entry is queryable with the same tools. There is one record of what happened, not two records that need to be merged.
Why a unified trail matters
A unified audit trail is not just a convenience for compliance teams. It changes how the operations team interacts with the system day to day. When a team lead reviews the morning's activity, they see one timeline. AI suggestions and human decisions are interleaved naturally. There is no need to switch between dashboards or cross-reference logs.
It also changes how problems are diagnosed. When a delivery was late, the trail shows exactly what happened: the AI proposed a reallocation at 09:15, the planner approved it at 09:18, the allocation was applied at 09:19, and the delay was detected at 11:40. The entire chain of causation is in one place.
The design constraint
When the AI's trail is separate from the human's trail, neither trusts the other. A unified audit record eliminates the reconciliation problem entirely. One trail, one schema, one source of truth. This is what makes AI actions reviewable, explainable, and auditable without additional tooling.
This is part 6 of 7 in the AI Trust Architecture series. Previous: Atomicity. Next: Human Accountability.