Agent Product Manager

Amigo•New York, NY

10d

About The Position

Amigo partners with healthcare organizations to deploy robust AI infrastructure that directly serves patients and providers. Our agents handle clinical workflows and patient engagement across the entire journey: pre-visit intake, care navigation, post-visit care plans, patient monitoring, and more. We own outcomes, not just delivery. For our customers, we're responsible for agent performance: clinical safety, continuous improvement, measurable patient outcomes. Agents operate autonomously within bounded clinical domains, with clear scope and handoff protocols. That scope expands as we validate performance across populations. We're backed by Tier 1 investors like General Catalyst, GSV Ventures, SVA, and CohoVC. Our work is validated with leading academic medical institutions. Our agents have reached 3M+ patient encounters and are on track to 10x this year. As an Agent PM at Amigo, you'll own agent quality and outcomes for your assigned customers. You'll lead domain expertise capture -- running structured interviews with clinicians to understand their reasoning patterns -- and translate that expertise into agent design decisions. You'll define what agents should and shouldn't do (Operational Patient Domains), design the evaluation strategy that proves they're working, and own the metrics that customers judge us by. This role requires deep judgment about where to draw boundaries in clinical AI.

Requirements

2-4 years of experience in product management, technical program management, or a domain expert role at a technology company
Experience working directly with domain experts (clinicians, lawyers, financial advisors) to extract and codify specialized knowledge
Strong analytical skills -- you can look at conversation metrics and distinguish signal from noise, identify systemic patterns vs. one-off issues, and prioritize by patient impact
Understanding of AI/ML systems at a conceptual level -- you don't need to write code, but you understand how LLMs work, why they sometimes fail, and what makes prompt engineering different from traditional software
Experience defining product scope and boundaries -- you've made "what's in / what's out" decisions and can defend them to stakeholders
Ability to communicate technical concepts to clinical audiences and clinical requirements to engineering teams
Track record of owning outcomes, not just activities -- you measure success by patient impact, not by features shipped
Project management discipline -- you manage multiple workstreams, track dependencies, and surface blockers before they become crises

Nice To Haves

Healthcare domain experience: clinical operations, health tech, digital health, or pharma
Experience with conversation design, chatbot products, or voice AI
Background in quality assurance, evaluation design, or measurement methodology
Familiarity with clinical workflows, EHR systems, or healthcare compliance (HIPAA)
Experience managing enterprise customer relationships
SQL or data analysis skills for deeper metric exploration

Responsibilities

Lead structured domain expertise interviews with clinical teams -- capturing reasoning patterns, decision frameworks, edge cases, and escalation criteria that become agent design inputs
Define Operational Patient Domains (OPDs): explicit scope boundaries for what agents handle vs. what gets escalated to humans -- with documented inclusions, exclusions, capabilities, and handoff protocols
Design agent behaviors and communication patterns: how agents greet patients, handle sensitive topics, manage frustration, respect clinical boundaries, and maintain appropriate tone across diverse populations
Own the evaluation strategy for your customers: define what metrics capture real quality (not gaming-prone proxies), set thresholds, design test scenarios that cover edge cases, and interpret results
Manage the agent improvement cycle: analyze conversation data to find systemic patterns, prioritize fixes based on patient impact, coordinate implementation with engineers, and verify outcomes
Own customer relationships independently -- running regular syncs, managing expectations during incidents, presenting performance reviews, and aligning on roadmap priorities
Design dynamic behavior specifications: define trigger conditions, instructions, and success criteria for contextual agent behaviors (appointment reminders, medication adherence, care plan follow-ups)
Coordinate cross-functional work across engineering, clinical affairs, and customer success to drive deployment milestones
Build and maintain the customer knowledge base: agent scope documents, clinical protocols, known limitations, and improvement history