Which TWO settings must be configured to ensure a Power Automate flow properly logs AI usage for audit purposes?
Trap 1: Add a 'Compose' action after every AI step
Compose actions increase clutter but don't provide centralized logging.
Trap 2: Change the Environment region
Region settings do not affect logging.
Trap 3: Set the flow to 'High Priority'
Flow priority governs run-order queuing under load, not audit logging. AI usage auditing requires enabling the relevant connector or environment-level logging settings and ensuring the flow's run history is retained; priority changes nothing about what gets recorded for compliance review.
- A
Activate AI Builder audit logs in the Power Platform Admin Center
Enabling AI Builder audit logs in the Power Platform Admin Center records AI model usage events, including which flows invoked which models and when. This satisfies the stem's audit-logging constraint by surfacing AI usage in Microsoft Purview audit search, where compliance teams can retrieve it alongside other tenant activity.
- B
Add a 'Compose' action after every AI step
Why it fails: Compose actions increase clutter but don't provide centralized logging.
- C
Change the Environment region
Why it fails: Region settings do not affect logging.
- D
Enable 'Tracked properties' in the AI action settings
Enabling tracked properties on the AI action exposes its inputs and outputs in the run history, which is what allows audit records to capture prompt and response metadata. Without it, the flow runs but leaves no auditable trace of AI usage.
- E
Set the flow to 'High Priority'
Why it fails: Flow priority governs run-order queuing under load, not audit logging. AI usage auditing requires enabling the relevant connector or environment-level logging settings and ensuring the flow's run history is retained; priority changes nothing about what gets recorded for compliance review.