Hi Community,
Our organization is exploring the rollout of autonomous workers using AI Agent Studio, and we are currently designing our internal governance model.
I am curious to hear how other teams are approaching this. Specifically:
- AI Guardian Configuration: Are most of you relying on the native ServiceNow SLM guardrails for prompt injection and offensiveness filtering, or are you utilizing custom external providers?
- Role & Privilege Design: How are you managing the 'Deny-by-Default' security posture? Are you creating dedicated integration service accounts per worker agent, or mapping them directly to user runtime criteria?
- HITL (Human-in-the-Loop): What thresholds or risk tiers are you using to determine when an agent can act autonomously versus when it must route to a human for approval?
Would love to hear any lessons learned or framework strategies from those who have deployed agents into production!
Thanks in advance!
