HubAIOps
← All use cases

Operations

Permissions & seats

AI only sees what the permission model exposes. Different roles see different data, so outputs change by who asks, and missing paid seats block agents entirely.

What breaks

  • AI insights that differ between sales and marketing
  • Summaries missing notes on restricted objects
  • Team members without the seats agents require

What we check

  • An admin can configure agents

    Do nowAuto-checked by scan

    Without the right role the setup stalls halfway, or gets done by whoever happens to have access.

    First step: Grant Breeze Agent Editor (or Super Admin) to the named AI owner.

  • Team has the required paid seats

    Do now

    Agents are unavailable to unseated users, so adoption fails for reasons that look like 'the AI doesn't work'.

    First step: Compare users against seat assignments and assign Pro+ seats to everyone in the pilot.

  • Permissions are documented and consistent

    Do now

    AI outputs change by role when visibility differs, which reads as inconsistency and kills trust.

    First step: Document permission sets per team and align object visibility for the records AI summarizes.

Related case studies