• AI visibility

Guide · AI search

Can AI search help a healthcare startup compete with established health brands?

It can, but only for startups whose clinical and financial results are public, independently confirmed and easy for an assistant to find. Employers and health plans now tie fees to outcomes, and patients increasingly ask AI tools about their conditions, so proof decides both who gets bought and who gets used. AI assistants rarely name young companies in open-ended questions, so a startup has to earn its place; no study yet shows this lifts a healthcare startup’s sales.

By Underneath Team Guide Updated

The short version

  1. Purchasers pay for results now: in the Peterson Health Technology Institute’s 2026 survey (opens in a new tab) of 321 digital health buyers, 68% of employers and 55% of health plans use performance-based contracts, and 85% of those contracts tie at least a quarter of fees to performance.
  2. Winning a contract is only half the sale: 47% of purchasers told PHTI that fewer than a quarter of eligible members enroll, and poor member engagement is the leading reason they switch vendors.
  3. Money is flowing, but unevenly: US digital health startups raised $14.2 billion in 2025 across 482 deals, according to Rock Health data reported by Healthcare Dive (opens in a new tab), while 35% of deals were flat or down rounds, Fierce Healthcare (opens in a new tab) reports.
  4. Patients have moved to AI: Rock Health found use of AI chatbots for health information doubled from 16% to 32% between 2024 and 2025, as Health Populi (opens in a new tab) reports, and KFF (opens in a new tab) also puts adult use of AI for health information at 32%.
  5. Newness is a handicap: in a study of 112 Product Hunt startups (opens in a new tab), ChatGPT recognized them by name 99.4% of the time but surfaced them in only 3.32% of discovery-style questions.

Who does a healthcare startup have to convince, and what is a customer worth?

Usually three groups: a purchaser who signs, the people who must enroll, and investors who fund the gap.

Purchasers. Most digital health startups sell through employers, health plans or health systems. Spending is holding up: in PHTI’s 2026 survey, 56% of purchasers plan to keep digital health spending at current levels over the next 12 months and 38% plan to increase it. But buyers are cutting the number of vendors. A Solera Health survey cited by D Magazine (opens in a new tab) found 42% of employers manage eight or more digital health vendors, and 90% of those spend more than $1 million a year administering them. Startups compete for fewer slots. Our guide on how healthtech vendors get named by employers and plans covers that purchaser sale in depth.

Members and patients. A contract pays off only when people use the service. A Baylor Scott & White executive told D Magazine that for many employer-purchased products, “the average usage rate is about two or three percent.” With performance-based fees, low enrollment can mean lost revenue as well as a lost renewal. Health plans face their own enrollment challenge, covered in how health insurers win members through AI.

Investors. Rock Health’s 2025 data show a split market. AI-enabled companies took 54% of funding, megadeals over $100 million took 42%, and average deal size rose to $29.3 million. Meanwhile, digital health M&A rose to 195 deals, some of them survival moves: Fierce reports that Thirty Madison’s valuation reportedly fell from $1 billion to $500 million in its sale. Biotechs courting partners and investors face a related search, covered in how biotech firms find partners through AI.

What a won customer can be worth shows in the sector’s public companies. Hinge Health, one of five digital health companies that went public in 2025, reported full-year 2025 revenue of $587.9 million, with 2,830 clients and 25 million contracted lives. Its value lies in contracts that put a service in front of millions of eligible people.

How far has AI entered the way patients and health buyers research options?

Deeply for patients, documented by surveys; for purchasers, AI use for vendor research is not yet measured.

Patients and members. Beyond the 32% figures, Health Populi reports that ChatGPT was used by 23% of Rock Health’s health information seekers and Gemini by 15%. Among people using AI for health, 59% explored treatment options based on a diagnosis and 55% researched prescription drugs or side effects. These users are also the digital health market: 84% of AI users had used an app or virtual care program in the past year, against 42% of non-users. OpenAI says over 230 million people a week (opens in a new tab) ask ChatGPT health and wellness questions.

Purchasers. PHTI found AI widely adopted inside health plans and health systems, mostly for administrative work such as clinical documentation (71% of health systems report some deployment). That measures AI in operations, not in vendor research. No survey yet tells us how often a benefits leader asks ChatGPT which virtual physical therapy or diabetes program to shortlist.

The live article on healthcare software and health system deals covers vendors selling IT into hospitals. This article is about the startup that has to win a purchaser and then the patient. Platforms selling virtual care to health systems and plans have their own guide on winning virtual care contracts and visits.

What do purchasers and patients ask AI about healthcare startups?

Category, alternatives, evidence and legitimacy questions. We wrote these examples ourselves; none were collected from real users.

Who asksIllustrative question
Benefits leader“Which virtual physical therapy programs have independent evidence of cost savings?”
Health plan“What are the alternatives to Omada for diabetes management?”
Benefits consultant“Which digital mental health vendors offer performance guarantees?”
Eligible member“Is the back pain program my employer offers any good?”
Patient“Are online menopause clinics legitimate, and how do they compare?”
Investor“Which women’s health startups have published clinical outcomes?”

A single question can trigger several searches. Google says AI Overviews and AI Mode may use a “query fan-out” technique (opens in a new tab), and OpenAI says ChatGPT search rewrites a prompt (opens in a new tab) into targeted queries. For “is this program any good,” we would expect those searches to reach for reviews, evaluations and news. Our study of how AI judges whether a business is legitimate looks at that kind of question directly.

Startups must keep their own content within limits: general information, no diagnosis, no personal medical advice, and claims that match the evidence.

How does an AI mention become a contract or an enrollment?

Through two paths: purchaser shortlists that lead to contracts, and member searches that lead to enrollment.

As we read the evidence, the two paths look like this:

Purchaser path.

  1. A benefits leader, consultant or plan executive asks about a category or alternatives to an incumbent.
  2. The answer names vendors and cites evaluations, news and company pages.
  3. The buyer requests evidence; independent reviews like PHTI’s matter here, because PHTI says digital health companies’ own return-on-investment estimates “vary in methodological rigor and reliability.”
  4. A contract follows, increasingly with fees at risk on outcomes.

Member path.

  1. An eligible employee or plan member has a condition and asks an assistant about options, or about the program in their benefits.
  2. If the answer describes your service accurately and points to it, enrollment becomes more likely, we infer.
  3. Enrollment and engagement drive the outcomes that performance-based fees and renewals depend on.

The second path is the one most startups overlook. A contract covering thousands of eligible employees is worth little if an AI answer tells them a competitor is the better-known option, or gets your eligibility rules wrong. Neither path is easy to trace in analytics; we cover the measurement problem in how to prove GEO caused sales.

Why do AI assistants name the established health brand and skip the startup?

Selection rules are undisclosed; studies show new companies are rarely surfaced and independent, institutional sources dominate health answers.

Documented by the platforms. Google and OpenAI confirm their AI answers search the web and link sources. Neither discloses why one health company is named over another.

Observed in studies.

  • The Product Hunt study found the gap between being known and being found: Perplexity recognized startups by name 94.3% of the time but surfaced them in 8.29% of discovery questions. The study found referring links and community presence related to Perplexity visibility.
  • In ChatGPT’s answers to consumer health questions, Jacques and colleagues (opens in a new tab) found 75.7% of cited sources came from institutions such as medical centers, government agencies and Wikipedia. Commercial platforms that were cited usually stated a medical review (71.1%). We summarize that study in what commercial health sites cited by ChatGPT have in common.
  • In our study of brand entities, independent coverage was the strongest predictor we measured: each tenfold increase in independent sites naming a brand in the cited pages went with 4.7 times the odds of being recommended.

What health buyers check. Evidence first. PHTI reported in 2023 that 80% of digital health products lack clinical evidence. Its later work, reviewed by the Society of Actuaries (opens in a new tab), found digital diabetes solutions generally increase spending while virtual musculoskeletal solutions can replace in-person care at lower cost. Independent verdicts like these are public, specific and citable.

Our inference. Established brands enjoy years of coverage, reviews and citations. A startup closes that gap by giving the assistant what a skeptical buyer wants: published outcomes, independent evaluations, named clinicians, and coverage by outlets with no stake in the sale. No study has yet tested this for healthcare startups specifically.

What does it cost a healthcare startup to be invisible in AI answers?

Mostly lost shortlists and weak enrollment, with funding consequences; no study has put a number on it.

  • Shortlists are shrinking. With employers consolidating vendors, a startup not named in early research may not make the reduced list.
  • Enrollment pays the bills. If fewer than a quarter of eligible members enroll for nearly half of purchasers, a startup that members cannot find or understand in AI answers loses performance fees and renewals, we infer.
  • Investors notice traction. In a market where 35% of deals were flat or down rounds, weak demand shows up in the next raise.
  • Misinformation spreads. Wrong pricing, coverage or eligibility details can appear in answers; see fixing wrong brand information in AI answers.

Which GEO work helps a young health company most?

Public evidence, independent validation and accurate, consistent facts that both purchasers and patients can find.

  1. One clear identity. State what you treat, for whom, through which purchasers, and in which states, consistently across your site, partner pages, benefits marketplaces and company databases.
  2. An evidence page. Publish peer-reviewed results, study designs and independent evaluations in plain language, with links. Avoid unsupported savings claims; purchasers and regulators read them too.
  3. Visible clinical oversight. Name your medical leadership and state how content is reviewed; the cited commercial health sites in the Jacques study showed this signal.
  4. Independent coverage. Trade press, peer-reviewed papers, evaluations and conference talks give assistants third-party confirmation. The method is in how brands build authority for AI search; how small brands get recommended by AI covers the startup case.
  5. Member-facing explainers. Write accurate pages on the conditions you treat, how to enroll through an employer or plan, and what it costs, without medical advice.
  6. Plan for the newness gap. Assistants often miss recent launches; see why ChatGPT misses new products. Then track purchaser and patient questions across ChatGPT, Gemini, Perplexity, Claude, Copilot and Google’s AI features, repeatedly.

Planted reviews and inflated outcomes are a legal and reputational hazard in health; see legitimate GEO versus manipulation. Broader startup lessons are in how AI startups win customers from AI search.

How solid is the case that AI visibility builds healthcare startup demand?

The supporting trends are well documented; the direct link from AI visibility to contracts or enrollment is not.

  • No purchaser data. No survey measures how often benefits leaders use AI assistants to shortlist digital health vendors.
  • The startup study is general. The Product Hunt sample covers technology products, not healthcare companies.
  • Health citation research covers medical questions. Purchaser questions may draw on different sources.
  • Several sources have interests. Solera sells navigation services; Hinge reports its own metrics; Rock Health invests in digital health.
  • No one has tied AI answers to signed health contracts. The wider evidence on revenue is thin as well; our review of whether AI visibility drives business results sets out what is known.

What should a healthcare startup do first to show up in AI answers?

Ask what your purchasers and patients ask, see who is named, then publish the evidence they cannot find.

List the questions a benefits leader, a plan executive, an eligible member and a patient would ask about your category, your competitors and your company by name. Pose each one to more than one assistant, and on more than one day, because a single answer can flatter or overlook a young company. Record which companies appear, which evaluations and articles are cited, and whether your outcomes, coverage and eligibility come back correctly. For most young health companies, the gap is evidence that exists but is not public, or is public only in a sales deck.

When you are ready, ask us for a startup AI visibility check. It shows where AI answers place you against established brands for purchaser and patient questions, which sources they trust, and which missing evidence is most likely costing you shortlists, member enrollment and renewals. How a young health company’s evidence, coverage and listings are then built up, step by step, is set out on our generative engine optimization service page.

Frequently asked questions

Do employers use ChatGPT to choose digital health vendors?

No survey measures it yet. PHTI’s 2026 survey shows purchasers focus on outcomes, with most employers using performance-based contracts.

Why does ChatGPT recommend the big brand instead of us?

Big brands have more independent coverage, which our brand study found was the strongest predictor of being recommended. New products are also often missing from answers.

Should we publish our clinical outcomes?

Yes, where you can do so accurately. Independent and peer-reviewed results are what both purchasers and AI answers can cite.

Can AI visibility improve member enrollment?

Plausibly, by helping eligible members understand and trust the program, but it is untested. Enrollment below a quarter of eligible members is common, PHTI found.

Is GEO worth it before product-market fit?

Usually not as a priority. It pays most once you have evidence to show and a purchaser channel to support.

Sources

Free strategy call

Some questions are easier to answer about your own business.

Bring the one that matters most. On a free 30-minute call we’ll take a first look at it and send you a short written read afterward.

Ask about my businessHow we work