AI Assistant

Approved Context

Active
Drafts
Current workspace: AI Assistant
Page usage manualManual

Approved Context guide

Use AI Assistant pages to draft controlled narratives from approved project evidence.

What it provides

  • Context builder for selected statistics, findings, and report metadata.
  • Narrative drafts with provenance and review boundaries.
  • Audit trail for provider, prompt, token, and context history.

How to use it

  1. 1. Inspect approved AI context used for reports.
  2. 2. Confirm processed statistics and findings are available.
  3. 3. Generate a draft from approved project context.
  4. 4. Review the AI output before using it in a report.

Outcome: AI output remains traceable, reviewable, and grounded in approved engineering evidence.

Approved drafts

0

Eligible report context

Attached

0

Used in generated reports

Requests

0

Provider calls

Context hash

-

Traceable input context

Approved Engineering Context

Approved context is built from project metadata, technical data, processed signal statistics, rule findings, report template intent, and review decisions.

Context Rules

Narrative generation should use only approved engineering context, never unrestricted raw logs.

Allowed

Project metadata, active technical files, signal statistics, rule findings, report template intent.

Excluded

Unbounded raw logs, unapproved DBC/ARXML/DTC/EOL files, private credentials, unrelated tenant data.

Trace

Provider, model, prompt version, token counts, context hash, review status, and report attachment.

Narrative Drafts

DraftProfileStatusReviewedAction
Loading...