Primary source: Menlo Ventures, 2026: The State of AI in Healthcare, November 16, 2026
On November 16, 2026, Menlo Ventures published its healthcare AI report. Spend grew 2.5x to $3.7 billion. The revenue-cycle line inside that total grew the slowest of the provider categories Menlo compares: coding and billing at 1.3x, against 1.6x for scribes, 4.5x for patient engagement and front office, and 5.5x for clinical copilots. In the same survey, 52% of payers rank AI-enabled provider coding among their top three cost drivers. Menlo is an investor. It marks portfolio companies with an asterisk.
Coding and billing growth in Menlo's provider comparison. Clinical copilots grew 5.5x to $550 million. Patient engagement and front office grew 4.5x to $450 million. Scribes grew 1.6x. These are Menlo's estimates, not an audited revenue-cycle market.
Menlo partnered with Morning Consult. In September 2026 the firm surveyed 550 U.S. executives at director level and above who have participated in buying an AI solution: 150 technology decision-makers at provider organizations, 100 senior executives at health plans and benefits administrators, and 300 executives at pharmaceutical and biotech companies. The dollar figures below are Menlo's estimates from that work. They are not a claims clearinghouse extract.
| What Menlo published | Figure | What it is |
|---|---|---|
| Healthcare AI spend | $3.7 billion, up 2.5x | Largest vertical. Menlo says it is more than double legal, and more than the entire vertical AI market a year ago ($3.5 billion, from Menlo's enterprise report). |
| Providers | 77% of spend, $2.9 billion | Up nearly 2x from $1.5 billion last year. |
| Payer external healthcare AI | $150 million, about 4% | Menlo says payers are spending 19x less than providers. |
| Coding and billing | 1.3x | Slowest provider category in the comparison Menlo prints beside scribes, copilots, and patient engagement. |
| Scribes | 1.6x; market $950 million | Startups led 67% of a $600 million market in 2025. Incumbents have 46% of $950 million in 2026. |
| New scribe dollars | 68% | The key findings say Microsoft and Epic took 68% of new scribing dollars. The scribing section says incumbent add-ons, led by Microsoft Dragon Copilot and Epic's Art, won 68% of new scribe dollars and grew 120%. AI-native scribes grew 28% and lost 13 points of share. |
| Clinical copilots | 5.5x to $550 million | OpenEvidence, marked as a Menlo portfolio company, is estimated at 64% of that revenue. |
| Patient engagement and front office | 4.5x to $450 million | The key-findings bullet compares this pair with coding and billing at 1.3x. The body adds the 1.6x scribe comparison. |
| Where buyers want to purchase | 71% coding and billing; 75% chart review | Share of providers who would ideally prefer those from their EHR. Menlo says the preference flips toward a third party for prior authorization (55%) and patient navigation (51%). |
| Implemented healthcare-specific AI | 42%, from 22% | Health systems 52%, outpatient providers 34%, life sciences 31%, payers 23%. |
Menlo's payer respondents are not describing their own AI budget. They are describing what landed on the claim. Forty-three percent of payers, and 57% of large payers, say provider documentation and coding AI raised their medical cost trend in 2026, up from 23% in 2025 and ahead of the 35% who, in 2025, predicted that outcome for 2026.
Fifty-two percent now rank AI-enabled provider coding among their top three cost drivers, behind only specialty drugs and provider rate increases. Over the next three years, 96% expect more codes per claim and 91% expect more prior authorization requests from providers. A year earlier, 80% to 83% expected increases. Menlo also says 85% of payers plan to invest in AI for prior authorization, utilization management, or claim validation, and it names Anterior, Autonomize, and Alaffia as startups building that response. Those three names do not carry the portfolio asterisk.
Prior authorization is now payers' number-one AI priority, at 82%, up from 59%. Menlo ties that to CMS's Interoperability and Prior Authorization Rule: 72-hour and seven-day decision deadlines this year, and electronic submission through standardized APIs by January 2027. Utilization management jumped the most, from 10% to 41%. Care management went from 46% to 61%. RevCycleAI covered the voluntary ONC and CMS testing sprint, and passing that test is not compliance, in a separate piece.
The spend gap is the other half of the same paragraph. Payer external spend on healthcare AI is $150 million. Menlo also reports that UnitedHealth Group says it will invest nearly $1.5 billion in AI this year. That $1.5 billion is UnitedHealth's statement, as Menlo relays it. It is not Menlo's measure of external payer software spend, and the report does not say how much of it is claim validation.
Seventy-one percent of providers would ideally prefer coding and billing from their EHR, and 75% say the same for chart review. Every major EHR Menlo names now ships a scribe: Epic, Oracle, athenahealth, and Meditech. Buying also slowed. Average time from first meeting to a closed deal rose from seven months to 10 for health systems and from five months to 11 for outpatient providers. Menlo states the same change as 52% longer for health systems and 128% longer for outpatient providers. Payer procurement time decreased 12%.
The startups that won the first scribe wave are leaving pure note-taking, on Menlo's account. In January, Abridge partnered with Availity to develop real-time prior authorization from the clinical conversation. In March, Heidi, marked as a Menlo portfolio company, partnered with R1 to bring payer policies and authorization requirements into the visit. Those are examples. They are not revenue figures for coding and billing.
This is RevCycleAI's read of the report. It is not a sentence Menlo printed.
Coding and billing is still growing. It is growing slower than the other provider categories Menlo lines up next to it, in a year when Microsoft and Epic took the new scribe dollars and when 71% of providers say they would rather buy coding and billing from the record system they already run. If you sell an independent coding or billing tool into Epic, Oracle, athenahealth, or Meditech, the preference number is the sales problem. Distribution is doing the work that a feature comparison used to do.
The payer response is the other side of a cleaner note. Providers bought documentation and coding tools first. Payers now say that work is raising medical cost trend, and they expect more codes per claim and more prior auth. Their external AI spend is still about 4% of the healthcare AI market Menlo sized. The priorities they named — prior authorization, utilization management, claim validation — are the counter-move. A coding vendor that only helps the practice submit a richer claim is funding the plan's next purchase.
Menlo's earlier enterprise report, which RevCycleAI covered on November 10, put broad enterprise AI deployment spending at $176.8 billion and named this same $3.7 billion as healthcare vertical AI. This paper is the healthcare cut. It does not publish a coding-and-billing dollar total. Do not cite $3.7 billion as the revenue-cycle AI market.
The methodology line is “*Menlo Ventures portfolio company.” Names that carry the asterisk in the report include OpenEvidence, Heidi, Anthropic, Chai Discovery, Assort Health, Benchling, Grow Therapy, Nourish, Qualified Health, and Yuzu Health. The closing note says Menlo partners with Anthropic, Chai Discovery, OpenEvidence, Benchling, Assort Health, Solace, Nourish, and Grow Therapy. OpenAI is named as a general-purpose model provider and is not marked. The category growth rates and the payer shares are still Menlo's published figures. They are not an association census, and they are not audited revenue.
Coding and billing AI grew 1.3x, the slowest provider category in Menlo's comparison, and 52% of payers rank AI-enabled provider coding among their top three cost drivers. The counter-move in the same report is payer prior auth (82%), utilization management (41%, from 10%), and claim validation. Seventy-one percent of providers would rather buy coding and billing from their EHR.
If you run provider revenue cycle, assume the EHR keeps pressing into the scribe, chart review, and coding, and assume a richer note draws a payer response. If you invest in an independent coding or billing vendor, the question is whether it still wins where the buyer wanted the system of record, and whether it has anything to sell the plan that now calls provider coding a cost driver. Menlo does not print a coding-and-billing dollar total, and it does not say what share of the $150 million in payer external spend is claim validation.
Who is growing, who already owns the buy, and which figure is an investor's estimate.
Sources: Menlo Ventures, “2026: The State of AI in Healthcare,” Greg Yap, JP Sanday, Croom Beatty, Johnny Hu, Derek Xiao, and Sabrina Lu, November 16, 2026 (page opened). Morning Consult survey design, the September 2026 fielding, and the 550 / 150 / 100 / 300 split are the report's methodology. The asterisk legend is the report's. Related: RevCycleAI on Menlo's 2026 enterprise AI report and ONC and CMS electronic prior authorization testing. No quote in this piece is invented. Menlo does not disclose a coding-and-billing revenue total, and none is estimated here.