Georgi Pepelyankov
PROBLEM
Workspace owners can't see or export what users ask Super Agents in direct messages, or what the agents reply. Super Agent Activity only shows run metadata (trigger, status, tools used). Audit Logs don't include transcripts. So a growing share of AI-driven work has no audit trail.
For regulated businesses this blocks adoption. We run a licensed online operation under ISO 27001 and GDPR, with a regulator overseeing us. We have to show what our AI agents were asked, what they did, and what data they touched. Right now we can't, so we have to restrict Super Agent use.
REQUESTED CAPABILITY
  1. Full transcript logging: every prompt, response, tool call and tool output, across DMs, channels, tasks and automations, with timestamps, user and agent IDs.
  2. Owner/admin access: a central view where owners or a delegated role can search and filter by user, agent, date range and keyword.
  3. Export: CSV/JSON export, plus API access or a SIEM/webhook stream so logs can go to our own retention system.
  4. Retention controls: configurable retention periods and legal hold, plus deletion of a single user's data for GDPR erasure requests.
  5. Tamper evidence: logs can't be edited or deleted by the end users who created them.
  6. User notice: an optional banner telling users their agent conversations are logged, to support transparency obligations.
WHY IT MATTERS
  • It closes the main governance gap that stops regulated companies (finance, gambling, healthcare, legal) from rolling out Super Agents widely.
  • It supports ISO 27001 logging and monitoring controls, GDPR access and erasure requests, and internal AI governance.
  • Competing enterprise AI tools already offer conversation audit and export, so this matters for enterprise deals.
PRIORITY FOR US: High. It's a precondition for wider Super Agent rollout across our Workspace.