Source: Fierce Healthcare — 2026-08-10
Summary
Doximity CEO Jeff Tangney laid out plans to scale the company's ambient notetaking tool (Scribe) and clinical AI assistant and search product (Ask) as part of what he's calling Doximity's 2026 "AI investment year." Central to the pitch is PeerCheck, a physician-verification layer that routes AI-generated clinical answers through a network of more than 10,000 medical experts before those answers reach doctors using the platform. Doximity is positioning this human-review layer as its differentiator in a crowded field of hospital-facing clinical AI tools.
Key Takeaways
- PeerCheck adds a human verification step between an AI-generated clinical answer and the physician who receives it, drawing on a pool of more than 10,000 medical experts to review and refine outputs rather than shipping raw model answers directly.
- Doximity is scaling two products in parallel: Scribe (ambient documentation) and Ask (a clinical AI search/assistant tool), both aimed at embedding AI directly into physician workflow rather than as a standalone chatbot.
- Tangney frames 2026 explicitly as an investment year, signaling the company expects near-term spending on AI infrastructure and headcount to outpace immediate revenue return from these tools.
- The company is leaning on its existing base of practicing physician users as a built-in review and distribution network — a structural advantage a general-purpose AI vendor without a physician user base doesn't have.
- The strategic bet is that trustworthiness, via visible physician sign-off, matters more to hospital buyers than raw model capability alone, a different competitive angle than vendors racing purely on benchmark accuracy.
Reel Script
Hook: Would you trust a clinical answer from an AI model, or one that over 10,000 doctors have personally checked first? Doximity is betting hospitals want the second one badly enough to pay for it.
Core Concept: The core problem with AI-generated clinical answers isn't just accuracy in the abstract — it's that a doctor reading an AI-generated summary has no easy way to know whether to trust a specific answer in a specific case. Doximity's fix is PeerCheck: instead of shipping an AI model's output straight to the requesting physician, the platform routes it through a review layer staffed by more than 10,000 medical experts who check and refine the answer before it's delivered. Think of it as a peer-review pipeline bolted onto a language model rather than a raw model-only pipeline — the AI drafts, humans with relevant clinical expertise check, and only then does the answer reach the end user. It's a slower, more expensive path than pure automation, but it directly targets the trust gap that keeps physicians skeptical of AI-generated clinical content.
Hands-On: The concrete mechanism to visualize is the pipeline itself: a clinical question comes into Doximity's Ask tool, the AI model generates a draft answer, and that draft routes through the more-than-10,000-strong PeerCheck network of medical experts before the physician user ever sees it — a human-in-the-loop layer sitting between model output and clinical use. Doximity is running this alongside Scribe, its ambient notetaking tool, as the two flagship products it's scaling in what CEO Jeff Tangney calls the company's 2026 AI investment year — deliberately front-loading spend on both products before expecting the revenue to fully catch up.
Takeaway: A 10,000-plus physician review layer is a genuinely hard moat to replicate — you can't buy that network overnight, and it directly answers the "can I trust this AI answer" question hospitals actually ask. Whether it scales without becoming a bottleneck as usage grows is the open question worth watching.