Doximity pushes clinical AI deeper into hospital workflows
Doximity has moved its clinical AI tools into more than 190 health systems, a sharp rise from 165 organizations in the prior quarter. The company says the customer base now includes 10 of the nation’s top 20 hospitals, a sign that enterprise buyers are no longer treating AI as an experimental add-on. They are folding it into the daily mechanics of care delivery, where documentation burden, chart review, and patient-specific decision support shape almost every clinician interaction.
At NewYork-Presbyterian, the company’s AI suite is now embedded directly in the electronic health record workflow. Clinicians across NewYork-Presbyterian, Columbia University Irving Medical Center, and Weill Cornell Medicine can access Doximity Scribe and Doximity Ask without leaving the clinical environment. The deployment spans 10 hospitals and hundreds of locations, extending a relationship that already included video and phone visits since 2021. In practical terms, the move places AI where physicians work rather than asking them to copy and paste chart fragments into a separate search tool.
Patient context is becoming the differentiator
One of the most clinically consequential additions is Doximity’s patient context feature, which UCSD Health is among the first hospitals to adopt. The feature allows the AI system to ingest EHR charts, laboratory data, and patient history so that responses are grounded in the record rather than in a generic prompt. Jeff Tangney, Doximity’s chief executive, said the feature is already changing how doctors query the system: prompts have become much shorter because clinicians no longer need to retype chart details.
The logic is straightforward. When an AI tool can read the problem list, notes, and labs, it can surface the relevant clinical signal faster and with less transcription error. In a high-volume inpatient or ambulatory setting, that can reduce friction at the exact moment a clinician is trying to answer a medication question, reconcile a diagnosis, or confirm a guideline-based pathway. The result is not just convenience. It is a different workflow architecture, one that leans on record-aware summarization rather than manual data entry.
Reliability and governance now sit at the center of the pitch
Doximity’s expansion is unfolding as hospitals tighten AI governance around data security, liability, and vendor controls. Tangney argued that health systems are converging on approved applications with enterprise privacy and security protections, and that trend benefits vendors able to operate inside institutional guardrails. The company says it has more than 300 enterprise clients across telehealth and other products, with privacy agreements in place across that base, though AI review is not universal across those customers.
The safety argument is not abstract. Patient data entering an unsecured AI workflow can create exposure if information is retained, reused, or surfaced inappropriately. For hospitals, the appeal of a vetted platform lies in reducing that uncertainty while still offering clinicians access to documentation tools, clinical questioning, and patient-specific support. Doximity is positioning its suite around that need, describing itself less as a search interface and more as an integrated clinical layer spanning telehealth, note-taking, scheduling, evidence review, and bedside questions.
Medication safety and evidence quality are part of the product strategy
Clinical AI tools rise or fall on trust, and Doximity is trying to answer that challenge with a set of reliability features. The company recently launched DoseCheck, a deterministic verifier designed to support medication safety. It also embedded UpToDate content into Doximity Ask through a partnership with Wolters Kluwer Health. The company argues that broad web-scale training alone is not enough for bedside decision-making, especially when clinicians face questions that do not have a randomized controlled trial answer.
UpToDate brings a different form of authority into the system: curated expert guidance written by thousands of physician authors and used widely at the point of care. Doximity is betting that blending evidence-based content with AI-generated clinical assistance will make the output more usable to doctors who need something closer to a reference standard than to a generic text generator. In a field where accuracy errors can affect prescribing, triage, and escalation decisions, reliability is not a marketing flourish. It is the product.
Usage is growing alongside the company’s broader workflow business
Doximity said quarterly active workflow prescribers rose 34% year over year, a faster pace than in the previous quarter. Roughly half of the 1.8 million U.S. prescribers, or about 900,000 clinicians, used Doximity workflow products during the quarter. Those tools include telehealth, scheduling, digital faxing, and communication products that sit close to the daily workflow of physicians. The company sees that usage as the base layer for wider adoption of AI features and, eventually, deeper monetization.
The commercial structure matters because Doximity is not treating AI as a single product. On the health system side, it charges a nominal amount as part of a broader workflow suite. On the pharma side, the company sees a separate revenue engine in AI Search, a commercial product launched in late April and already purchased by 16 of the top 20 pharmaceutical companies, with three more in contracting. Bookings have reached eight figures, and the company expects that revenue to ramp gradually through calendar 2027.
Hospital adoption and commercial AI are moving on parallel tracks
The company is pursuing two very different markets with the same technology stack. Clinicians want documentation support, patient context, and safer answers at the bedside. Pharma clients want AI Search as a new channel for digital promotion and paid search. Doximity says that paid search has historically pulled in substantial marketing spend, and it plans to move toward auction-based purchasing to make the product behave more like conventional search advertising.
Financially, the quarter reflected that dual-track strategy. Revenue for the period ended September 30 reached $174.5 million, up 4% year over year and above expectations. Adjusted EBITDA came in at $89 million, or a 51% margin, even as AI spending rose more than sixfold year over year. Net income declined to $34.7 million from $62.1 million a year earlier, while adjusted earnings per share were 36 cents. Doximity also lifted its full-year revenue outlook to $678 million to $683 million.
The hospital-side story is still about adoption and trust. The unanswered question is how quickly health systems will standardize around a small number of approved AI tools, and whether those tools will be measured by documentation efficiency, clinician satisfaction, error reduction, or some combination of all three. For now, Doximity is pressing its case with broader EHR integration, tighter governance controls, and a clinical AI suite built to live inside the workflow rather than beside it.