Audit Logs
The AI Usage Audit shows how AI features are being used across your firm: which advisor ran which feature, for which household, and how much it consumed. It is the record you use to demonstrate that each household's AI Usage Policy was actually honored.
What Sterling Records — and What It Does Not
Every metered AI action settles into a usage ledger on Sterling's servers as it happens. Each entry records the advisor, the feature, the household, the timestamp, and the volume consumed.
No Prompts, No Responses
The audit deliberately stores no message content. Not the advisor's prompt, not the data sent to the model, not the text that came back. Keeping transcripts would mean holding client information at rest for no compliance benefit, so Sterling records that an action happened rather than what was said.
This is a change from earlier versions of Sterling, which stored full prompt and response transcripts alongside a per-message PII score. That log was retired and its stored content deleted. If you are looking for the message-comparison view, it no longer exists — and the questions it was meant to answer are answered better by the policy and usage records described below.
Why Audit Logs Matter
Financial services firms face increasing regulatory scrutiny around AI usage. This record gives you documentation for:
- Honoring client preferences: Show that households set to No Access had no AI processing
- Regulatory compliance: Document when AI was used, by whom, and on whose behalf
- User accountability: See which team members use which AI features, and how often
- Policy governance: Show who changed a household's AI policy and when
The Three Views
The dashboard is organized into three sections.
Usage by User and Action
Who used which AI feature and how much. Firm administrators see the whole firm; everyone else sees their own activity only.
Household Compliance
Available to firm administrators. Every household in the firm, listed with its current AI Usage Policy and its usage totals side by side. This is the view that does the compliance work:
- A No Access household showing zero usage is your proof that the setting was honored
- A No Access household showing any usage is flagged as a violation
- Usage that carries no household context is listed in its own bucket rather than being attributed to a client
The Report You Actually Need
If a client or examiner asks whether a household's data was ever processed by AI, this view answers it directly — policy and usage on the same row, for every household at once.
Policy Change History
Available to firm administrators. Every change to a household's AI Usage Policy, with the team member who made it, the date and time, and the previous and new values. Together with the usage view, this lets you establish which policy was in force on any given date.
Filtering
Narrow any view by date range, team member, household, or AI feature. Firm administrators can filter by any user; standard users are always scoped to their own activity.
Access Permissions
- Firm Administrators: All three views, across every user and household in the firm
- Standard Users: Their own usage only. The household compliance and policy history views are administrator-only
Accessing the Audit Log
- Click your profile icon in the top navigation
- Select Audit Log from the dropdown menu
- Choose a view and set your filters
- Review the results
Regular Reviews
Consider reviewing the household compliance view monthly. A violation flag is far cheaper to resolve the week it appears than during an examination.