Portfolio Intelligence Agent
Synthesizes many signals across your book and reasons about which accounts truly need intervention — not just threshold alerts.
Portfolio Intelligence Agent
What it does. Rather than firing an alert every time a single number crosses a line, this agent reasons across your whole portfolio like an experienced CSM would — weighing multiple signals together and deciding which accounts genuinely need intervention this week. It can also decide an account is fine and *decline* to act, which keeps your inbox honest.
Signals it reads. Health scores and their trend, the churn score, renewal pipeline and timing, expansion readiness, product usage, contacts and NPS, and how similar accounts have played out (via cross-account similarity).
What it produces. One context-aware recommendation per account that needs it, written to the AI Agent Inbox for your review. Each recommendation carries a reasoning trace and a proposed task, so you see *why* before you approve.
Autonomous vs on-request. Autonomous — it runs on a daily schedule — and you can also trigger it manually from the AI Agents admin page.
Your control. Enable/disable and configure it in Admin Console → AI Copilot & Agents → Agents. It respects a cooldown so it won't re-recommend the same account repeatedly. Recommendations are never applied without human approval.