AI is reshaping the doctor visit—just not how you think

Digital health startups raised $14.2B in 2025, up 35% year-over-year, with AI-enabled companies capturing 54% of that capital and commanding ~19% larger deal sizes. Zocdoc's survey of 1,186 U.S. adults and 1,000 providers finds 26% of patients have asked AI health questions, 85% of providers see more AI-informed patients, but >20% of patients hide AI use and 83% of providers report correcting AI information; 70% of patients still prefer a doctor over AI and average PCP wait times top 31 days.

Analysis

AI‑informed patients are creating a hidden friction tax on clinic throughput: clinicians spend nontrivial time unstitching AI outputs from clinical context, which mechanically reduces visits per clinician and raises per‑visit labor intensity. That creates a clear arbitrage for software that captures and monetizes the extra workflow (scribes, scheduling, pre‑visit triage) rather than consumer apps that merely answer questions. Expect margin transfer from low‑value consumer query endpoints to workflow incumbents who can embed billing capture and decision provenance.

For platform and ad incumbents, the threat is asymmetric: commoditization of query intent reduces search ad yield while enterprise capture of clinical workflows increases sticky, subscription‑style revenue. Cloud/compute providers and EHRs stand to gain from running and integrating large models into EMRs, but near‑term deployment will be constrained by malpractice/regulatory scrutiny and integration costs. A 6–18 month horizon is realistic for visible revenue lifts in enterprise contracts; consumer app valuations will reprice faster if regulatory audits expose accuracy issues.

Catalysts to watch: (1) payer guidance tying reimbursement to AI‑assisted documentation, (2) a wave of hospital RFPs for scribe/triage integration, and (3) any high‑profile adverse outcome that prompts stricter consumer AI rules. Reversal risks include rapid improvement in consumer AI accuracy or new low‑friction telehealth capacity that reduces the current access gap. Position sizing should favor deterministic workflow revenue capture with hedges against ad‑market cyclicality.

Key Decisions for Investors

• Buy ORCL 12‑18 month call spread (bull call spread ~10% ITM vs 25% ITM): Oracle/Cerner integration is best placed to monetize AI‑scribe and billing capture in hospitals. R/R: asymmetric — limited premium vs high probability of multi‑quarter SaaS revenue re‑rating if hospital RFPs accelerate; risk: long sales cycles and integration delays.
• Initiate a 6–12 month bearish put spread on GOOGL/GOOG (buy 15% OTM puts, sell 30% OTM) sizing 1–2% portfolio: hedge vs structural ad‑yield erosion as queries migrate to specialized health interfaces. R/R: collects upside if search monetization weakens; risk: Google offsets with alternative ad products or stronger cloud growth.
• Long TDOC stock or 9–12 month calls (moderate size): telehealth vendors can arbitrage the access gap and absorb triage volume funneling from consumer AI. R/R: direct beneficiary of higher teletriage demand and scheduling; risk: margin pressure and competition from integrated health systems.
• Pair trade — long NVDA (or AMZN/AWS exposure via calls) and short GOOGL (equal notional, 6–12 months): play infrastructure winner (model hosting, inference margins) versus search ad exposure. R/R: captures secular shift to model compute spend; risk: macro tech rally that lifts both legs.