Agent Fleet is the operator console for the AI SRE workforce. It shows every registered agent, what they are doing right now, how often they succeed, and how much trust they have earned over time. Use it to spot stalled agents, audit recent decisions, and tune autonomy without leaving the page.
Nova AI Ops ships 100 expert backend personas grouped into 12 teams: SRE Core, Cloud, DevOps Tools, Cybersecurity, AI Science, Specialists, Security, Data and ML, Business, plus seven meta-orchestrators (Task Planner, Agent Router, Conflict Mediator, and friends). Each team owns a slice of the operations surface area, so a Postgres deadlock is handled by a Postgres specialist, not a generic LLM.
Each agent card shows live state (idle, running, paused), task count over the last hour, success rate, p95 task latency, and the agent's current autonomy level. Click any card to drill into recent decisions and the decision bundle behind each one. The grid auto-refreshes every two seconds so you never act on stale data.
Every action an agent takes adjusts its trust score. Successful, verified remediations push it up. Rolled-back actions, ground-truth verification failures, and operator overrides push it down. Trust unlocks higher autonomy tiers (shadow, suggest, auto-rollback, fully auto). The tier history graph on each agent card shows when an agent moved tiers and why.
New to the fleet? Start at the roster grid and filter by your team. Click any agent to see its decision history. Use the autonomy dropdown to demote a noisy agent. Hit the kill switch in the top-right corner if anything looks off. Everything you do here is captured in the audit log automatically.
Subscribe to Nova AI Ops on YouTube for demos, tutorials, and feature deep-dives.
See every AI agent at a glance. Pause one. Promote another. Audit any decision back to the prompt and the response.