Healthcare products that cleared institutional adoption
Automated clinical decision rules across 100+ hospitals at Epic, an ML salary model on 35,000-physician data at Doximity. Symptom-checking triage and routing across four programs at Transcarent, oncology navigation for 30MM annual visitors at CancerCompass, automated contract analysis at Andwise. Each one required product systems that turned regulatory ambiguity, clinical risk, and trust gaps into something an institution would deploy.
The identity decision in the product's own words — the clinician picks which number the patient sees, with the office line preselected.
2013–2017
The identity infrastructure behind trusted telehealth
Mandate
Early product lead on Doximity Dialer. Owned the identity/verification layer, HIPAA compliance policy, and the EHR integration strategy that put verified calling inside clinical workflows.
Constraint
Physicians need to reach patients but face a structural bind: call from the hospital desk phone (geographically tethered) or call from their personal cell (exposes their personal number permanently — no boundary, no audit trail, safety risk). Most choose neither. The call doesn't happen. And when they do call from a cell phone, patients see 'Unknown Caller' and don't answer. Twilio required verifying access to the outbound phone number before letting us display it as caller ID — the legitimate path, but slower. We found a vendor that let us set any caller ID without verifying ownership. I knew that approach would never pass hospital IT security review. But I also knew hospital IT would likely never approve the product at all — it lets physicians make outbound calls using their office number from a personal cell phone, geographically liberating the number from the hospital phone system. No security team green-lights that on paper. The only path to institutional adoption was to get physicians using it first — build critical mass, and make approval a fait accompli rather than a request.
Decision
Took the shortcut — used a vendor that didn't require caller ID verification — because the legitimate path would never get approved. Hospital IT wouldn't pre-approve a product that lets physicians use their office number from a personal cell phone. The only way through was around: ship to physicians directly, build critical mass, and let institutional approval follow adoption instead of preceding it. Built the HIPAA compliance policy and Epic Haiku integration to give hospital IT a documented basis to clear the product once it was already in use.
Operating model
Designed and shipped HIPAA-compliant verified caller-ID and fallback workflows ensuring the clinician's office was the visible, trusted identity on all outbound calls. Authored the HIPAA compliance policy — policies, procedures, and the white paper that gave hospital security reviews a documented basis to clear the product. Partnered with legal and hospital IT to clear EHR integration across multiple systems. Led the Epic Haiku integration putting one-tap verified calling inside the chart the clinician was already in.
Outcome Established the identity layer Dialer scaled on. Nine years after I left Doximity, the product is still the #1 ranked telehealth platform by health system CIOs — five years running, beating Microsoft Teams and Zoom. Doximity now conducts 300,000+ voice and video visits per average weekday across 200+ health systems, reaching ~45% of U.S. physicians with paid enterprise licenses. Five direct competitors tried to build a physician-focused Dialer and failed. Pickup rates 3x higher than competitors — patients see the office number, not 'spam risk.' The architecture held.
Why it matters
When the call connects, a lab result gets communicated tonight instead of next week. A post-discharge follow-up reaches the patient before a complication sends them back to the ER. A physician who gets a biopsy result at 7pm can call from home — the patient sees the office number and answers, the personal boundary stays intact, and the diagnosis doesn't wait until morning. The first sign it was working wasn't a metric: medical residents I'd gone to undergrad with told me they were getting more sleep at night — they could call patients from home with the office number instead of staying at the hospital to use the desk phone. At Transcarent, the problem changed from trust to scale: how to make that same trust work across four clinical programs at once.
Led the product from prototype through physician adoption and subsequent institutional security clearance. Authored the HIPAA compliance policy. Partnered with legal and hospital IT to clear EHR integration once physicians were already using the product.
Built the identity infrastructure powering 300,000+ daily voice and video visits across 200+ health systems (per Doximity Q2 FY2026 earnings call, Nov 2025).
Put verified identity directly into the clinician's primary workflow via Epic Haiku.
Authored the HIPAA compliance posture that gave hospital security teams a documented basis to clear the product.
Drove iOS ratings from 3.7 to 4.8 stars through reliability and trust engineering.
Transcarent public app visual for care-navigation context.
2021–2022
Scaling specialty-care product strategy across four clinical programs
Mandate
Led product strategy and cross-functional delivery for value-based specialty-care and care-navigation programs. Managed product managers, architected the shared decision infrastructure, and coordinated across clinical, engineering, and commercial teams.
Constraint
High-acuity specialty cases (Surgery, Oncology, Behavioral Health) were stalling in exception queues, limiting the speed and scale of program deployment across employer partners. The organizational challenge was scaling across four clinical domains simultaneously — each with different clinical protocols, benefit structures, and partner integrations — without building four independent products.
Decision
Chose to build shared decision infrastructure rather than per-program products. Designed a unified, role-based automated ownership model that tied routing rules to clinical accountability, enabling four programs to scale on the same architecture. Made the tradeoff explicitly: slower to build, faster to scale.
Operating model
Designed unified member records and tied routing rules to clinical roles, ensuring ownership was explicit at every decision point in the care journey. Managed and mentored product managers across four programs. Coordinated across clinical, engineering, and commercial teams to align on routing logic, exception handling, and escalation protocols. Led the Cancer & Decision Support pod: shipped CancerCare v1, a symptom-checking triage tool with Healthwise content integration and an in-app second-opinion experience with automated handoff to Nurse Ally through Consumer Medical.
Outcome Four specialty-care programs launched on shared infrastructure. Routing-related delays reduced as a primary bottleneck. Operational variance between highest- and lowest-performing service lines narrowed, improving consistency at scale.
Why it matters
The shared decision infrastructure made ownership explicit at every decision point, which is what let four programs scale on the same architecture instead of four separate products. At Andwise, I tried to apply the same accountability pattern as a founder — and learned that building the product is necessary but not sufficient.
Managed and mentored product managers across four specialty-care programs; coordinated across clinical, engineering, and commercial teams; architected the shared decision infrastructure.
Four specialty-care programs launched on shared decision infrastructure.
Routing-related delays reduced as a primary bottleneck.
Operational variance between highest- and lowest-performing service lines narrowed.
Clinical decision-making transformed from fragmented, exception-based manual work into automated routing with human-in-the-loop escalation for high-risk cases.
MyChart — the EHR surface where the decision rules fire, and where institutional workflow architecture has to work.
2012–2013
Clinical decision systems adopted across 100+ hospital go-lives
Mandate
Implementation Engineer / Technical Services. Implemented clinical decision-support rules, quality-measurement workflows, and designed the escalation process adopted company-wide. Coached 100+ hospitals through go-lives as subject-matter expert.
Constraint
Each hospital go-live required hundreds of clinical decision rules, but there was no systematic escalation path when regulatory updates or critical bugs landed mid-deployment. During go-lives, I saw how close a misrouted alert could come to burying a critical lab result. Diverse hospital policies made per-organization customization the default, creating maintenance debt that would compound across 100+ implementations.
Decision
Built reusable rule configurations rather than per-hospital customization. Automated Best Practice Advisories that triaged alerts between binding and advisory based on policy constraints, and a company-wide escalation process that gave regulatory updates and critical bugs a defined owner and response time.
Operating model
Implemented clinical decision-support rules in Epic's rule engine that mapped hospital policy constraints into automated triage logic — reducing alert fatigue without losing safety checks. Drove adoption of the escalation process across Epic's internal teams. Supported named go-lives at Owensboro Health (hospital relocation) and UAMS, and coached 100+ hospitals as subject-matter expert.
Outcome 100+ hospital go-lives supported with configured clinical decision rules. The escalation process was adopted company-wide.
Why it matters
Origin of the escalation-design philosophy I still practice — defining default responses before the edge case arrives. It's also where I first saw that institutional workflow architecture, not feature quality, determines whether clinical software gets adopted — the lesson I carried to Doximity, where the problem was trust, not routing.
Coached 100+ hospitals through go-lives as subject-matter expert; designed and drove adoption of a company-wide escalation process adopted across the organization.
100+ hospital go-lives supported with implemented clinical decision rules.
Company-wide escalation process adopted for regulatory updates and critical bugs.
Alert fatigue reduced without losing safety checks through policy-constraint mapping.
Meaningful Use and quality-measurement workflows implemented at production scale.
CancerCompass archival public surface — the oncology navigation platform serving 30MM annual visitors.
2019–2021
Scaling oncology navigation for 30MM annual visitors
Mandate
Director of Digital Products & Strategy. Led relaunch and product strategy for an oncology information, navigation, and patient-engagement platform. Directed $2M in software development execution.
Constraint
An oncology information platform with 30MM annual visitors needed to translate clinical protocols into actionable steps for patients and families navigating cancer care.
Decision
Led relaunch with plain-language navigation flows, feedback loops with nursing teams to tune triage rules, and reliability reviews so content updates stayed aligned with clinical oversight.
Operating model
Directed product strategy and $2M in software development execution. Wrote the Director of Development job description and defined the tech-stack architecture for the platform's rebuild. Built feedback loops with nursing teams to continuously tune triage rules and reduce avoidable escalations.
Outcome Reduced bounce rate by 25% and increased chat conversions by 267% through interface and analytics updates. Directed $2M in software development across core platforms.
Why it matters
Translating clinical complexity into consumer-scale product — 30MM annual visitors navigating cancer care from a screen, not a phone call.
Directed $2M in software development execution; defined tech-stack architecture and hired the Director of Development for the platform rebuild.
Reduced bounce rate by 25% through interface and analytics updates.
Increased chat conversions by 267% through interface and analytics updates.
Directed $2M in software development execution across core platforms.
Translated clinical protocols into plain-language steps patients and families could follow at scale.
2022–2024
Scaling trusted financial guidance through automated review and accountable sign-off
Mandate
Co-Founder and Product Leader. Built the medical advisory board, review routing, escalation clocks, and accountable sign-off workflows that enabled scaling trusted financial guidance to physicians.
Constraint
Early-career physicians faced high-friction, high-stakes financial decisions (debt, contracts, home buying) without access to reliable, unbiased, and scalable guidance.
Decision
Led the development of automated clause-level analysis that flagged key contract terms for review, paired with medical advisory board oversight and escalation-clock routing to ensure accountable human sign-off.
Operating model
Convened a medical advisory board of over 50 physicians, designed review routing with escalation clocks, and implemented plain-language summaries with accountable sign-off.
Outcome Scaled to 1,200+ physician users and compressed standard recommendation turnaround from days to hours. Andwise wound down in 2024 — not because the product failed, but because investors wanted us to monetize fear. We chose to shut down rather than become what we set out to protect physicians from.
Why it matters
The escalation-clock pattern is the same authorization-state architecture that appears in the AI governance work. The wind-down taught me that the business model has to be designed before the product, not after — or you'll be forced to choose between the mission and the money.
Co-founded and led the company; built and convened a 50+ physician advisory board, raised $240K, and managed the product team through wind-down.
Compressed recommendation turnaround from days to hours by automating review routing and escalation management.
Scaled to 1,200+ physician users and a 700-member community through high-fidelity, trusted guidance.
Enabled high-stakes advisory work at scale by encoding fiduciary and compliance review into the product core.
Andwise wound down in 2024 — investors wanted us to monetize fear (tiered plans, referral fees, urgency tactics). We chose to shut down rather than become a fear-based funnel.
If getting AI through clinical adoption is your bottleneck, email me.
