Principal Product Manager, Agentic AI

About the role

Suits someone who has taken an agent product from early research to enterprise rollout and is comfortable making calls about pay and legal risk. Much of the work happens before launch, so direction gets set rather than inherited.

At a glance

What you will do

• Decide what each agent may do, which tools it can call, when it must escalate and what a manager sees
• Spend time in customer operations and rank which workflows an agent should and should not take on
• Define autonomy limits, override paths, logging and audit trails for actions that affect pay and compliance
• Stay on the build through implementation, review it against the spec, join bug triage and hold releases
• Write and maintain evaluation sets, review failure traces and set the pass criteria for shipping

What they look for

• Has shipped agentic or AI products through specification, prototyping, evaluation and enterprise deployment
• Can turn messy operational research into clear specs and defend the priorities behind them
• Is hands-on with implementation and willing to use the product before customers do
• Has worked on products where pay, legal compliance or similar stakes were at play

Nice to have

• Has written functional requirements documents as a maintained source of truth
• Has handled packaging, metering or pricing questions for software that does work people used to do