Product Manager

Role Overview

Journey ownership

Own discovery → evaluation → booking → pre-visit confidence. Remove friction and uncertainty with transparent information and safe guidance.

AI-assisted UX

Ship AI features (e.g., curated answers, provider matching, safety nudges) using retrieval and guardrails; instrument quality and costs.

Trust & safety

Embed our policies (no medical advice, CPOM awareness, HIPAA boundary) into UX. Balance speed with safety and compliance.

Experimentation

Hypothesis → design → A/B test → causal analysis → rollout. Maintain a clear metric tree and avoid local maxima.

Cross-functional

Partner with Design, Eng, Data, Provider Ops, and Legal. Communicate trade-offs in writing (PRDs, briefs, postmortems).

• Higher qualified bookings and patient satisfaction (NPS/CSAT) with measurable uplift.
• Reduced time-to-answer and drop-off via AI-curated guidance and safer defaults.
• Improved provider responsiveness through smarter routing and expectation setting.
• Clearer signals of quality (verified reviews, profile completeness, Service Score™ inputs).
• Fewer escalations through proactive safety nudges and "right next step" UX.

Qualified booking rate

Definition: Bookings that meet geo/treatment/availability criteria and pass fraud checks.

Guardrails: No increase in cancel rates; maintain response SLAs.

Decision latency

Definition: Time from first visit to inquiry/booking.

Guardrails: No rise in misroutes or safety escalations.

AI answer quality

Definition: Offline eval + human review pass rate for curated answers/guides.

Guardrails: No medical advice; privacy boundary respected.

Patient satisfaction

Definition: Post-booking CSAT/NPS for patient journey.

Guardrails: Bias/representation checks; access equity.

Cost efficiency

Definition: Revenue or bookings per AI/infra dollar.

Guardrails: Quality and safety metrics stay green.

• First 30 days
  • Map current patient journeys, metrics, and pain points; review safety & privacy constraints.
  • Ship a small, high-signal improvement (copy, ordering, default) with an A/B.
• Days 31–90
  • Own a multi-step initiative (e.g., AI-curated evaluation guide or smarter provider matching).
  • Establish a metric tree and an experiment backlog; publish a public (internal) roadmap.

• Must-have
  • 4+ years as a Product Manager shipping user-facing experiences at scale.
  • Strong experimentation skills (A/B, metrics, causality basics); clear PRDs and crisp writing.
  • AI-aware product sense: knows when to use retrieval, prompts, and guardrails vs. classic UX/IR.
  • Comfort with data: funnels, cohorts, retention; partner tightly with Data/Eng.
  • Excellent collaboration across Design, Eng, Ops, and Legal.
• Nice-to-have
  • Healthcare marketplace, UGC/reviews, or trust & safety experience.
  • Familiarity with privacy concepts (consent, minimization, retention), HIPAA boundary, and CPOM sensitivity.
  • Light technical depth (e.g., understands APIs, search, RAG at a conceptual level).

Outcomes over hype

Use AI where it measurably helps patients decide and book safely. When a simpler UX or search improvement beats AI, choose the simpler path.

• Define problems that AI can realistically improve; avoid overreach into medical advice.
• Work with Eng to set quality bars (offline eval sets, human review, red-team prompts).
• Design guardrails: privacy boundary, prompt injection resistance, safe defaults and disclosures.
• Instrument cost/latency/quality; maintain dashboards and review weekly.
• Document decisions and model/dataset provenance; plan fallbacks and rollbacks.

No medical advice. Do not route PHI or consumer health data to non-approved endpoints. Prefer retrieval of vetted content and clear hand-offs to providers.