Principal Product Manager, AI Agents & Automation

The role

Every day, storage operators field calls and messages from people who want to rent a unit or need help with one: prices, availability, payments, gate access, move-ins and move-outs. Every missed call can be a lost rental, and many of these requests touch payments and facility access. You'll own the AI agents that handle this work, end to end: the voice and text experiences prospects and tenants use, the rules for what agents can do in our systems, and the quality bar that keeps every agent accurate, safe and affordable.

What you'll do

• Own the agents that serve prospects and tenants across voice and text, and the automations that take routine work off facility staff, from request to completion.
• Choose which workflows agents take on first, and which come next across Storable's products as model capabilities improve, working with Operations and Support.
• Set the rules for every action: what agents may do, what needs verification, confirmation or staff approval, and when to hand off to a person. Work with Legal and Compliance where requirements vary by state, such as collections and liens.
• Define what operators can configure themselves, including knowledge, policies, escalation rules and brand voice, so agents work for a single facility or a large portfolio without custom work.
• Define what good looks like for each workflow: success criteria, agent evaluations, quality thresholds and launch gates, tracked in production.
• Partner with the APIs & Developer Platform team on the shared APIs and MCP tools agents call.

What you bring

• Experience as a technical product manager shipping and operating an LLM-powered product that takes real actions for users, not just answers questions: the experience, the actions and how it's evaluated.
• Strong understanding of how agent systems work: models, context and retrieval, tool calling, orchestration and memory, and how each choice trades off quality, latency and cost. You can read a trace and explain what went wrong.
• A track record of defining what good looks like for nondeterministic products: success criteria, quality thresholds and launch gates that decide what ships.
• Experience setting the rules for agents that take real actions: verification, permissions, confirmation, approvals and handoff to people, including in regulated workflows.
• A record of leading ambiguous, cross-team work from idea to launch.
• Excellent writing. Your product and technical documents are clear enough that teams can decide and build from them.
• You build working prototypes yourself, including agents, with modern tools.

Nice to have

• Voice or contact center products: speech, telephony and real-time latency.
• Building on MCP or agent frameworks, or evaluating new models as they ship.
• Workflows involving payments, identity or collections, where a mistake costs money.

All applicants must be currently authorized to work full-time in the United States and able to work a hybrid schedule in our Seattle, WA office.