• AI visibility

Guide · AI search

How do telehealth companies get chosen when patients ask AI where to get care?

By being the virtual care service an assistant can verify when someone asks where to be seen: clear prices, the insurance you accept, the states you serve, licensed clinicians and a reputation that holds up on review sites. A third of US adults now ask AI chatbots about their health, and most who ask about physical health go on to contact a provider. The telehealth brands that are named, described accurately and trusted at that moment can win the visit; the rest pay more to be found.

By Underneath Team Guide Updated

This article is about how consumer telehealth companies are found and chosen. It is not medical advice, and nothing here recommends a diagnosis, treatment or provider. For software sold to hospitals and health systems, see our guide to healthcare software and AI search.

The short version

  1. In KFF’s March 2026 poll (opens in a new tab), 32% of US adults had used AI for health information or advice in the past year, and 58% of those who asked about physical health later followed up with a doctor or other provider.
  2. Access is part of the reason: 18% of those AI users said a major reason was that they had no regular provider or could not get an appointment, and 19% that they could not afford to see one.
  3. Acquisition is expensive: Hims & Hers (opens in a new tab) spent $262.2 million on marketing in the second quarter of 2026 on revenue of $753.2 million, while Teladoc’s BetterHelp (opens in a new tab) averaged 346,000 paying users, down 11% from a year earlier.
  4. Paid channels are gated: Google Ads (opens in a new tab) allows telemedicine providers to advertise only if they are verified by LegitScript’s Healthcare Merchant Certification Program.
  5. The assistant can itself be a competitor: Amazon’s Health AI (opens in a new tab) connects users to One Medical clinicians, with five free consultations for more than 30 common conditions for eligible Prime members and $29 pay-per-visit care after that.

Who chooses a telehealth service, and what is a patient worth?

Mostly younger, convenience-minded adults, many without a regular provider, paying by subscription, per visit or through insurance.

Telehealth became routine during the pandemic and then settled. National Health Interview Survey data reported by HealthDay (opens in a new tab) show the share of adults who used telemedicine in the past year fell from 37.0% in 2021 to 30.1% in 2022. That is still roughly three in ten adults, which makes virtual care a mainstream choice rather than a niche one. In-person practices compete for the same patients; see how clinics and hospitals win patients from AI.

Consumer telehealth brands earn money in three ways, and the value of a new patient depends on which:

ModelReal example (public figures)
Subscription careHims & Hers had 2.891 million subscribers at the end of Q2 2026, up 19%, and monthly revenue per average subscriber of $92
Therapy subscription and insuranceBetterHelp segment revenue was $212.6 million in Q2 2026, down 12%, as demand shifted toward insurance-covered care
Pay per visitAmazon One Medical charges $29 for a pay-per-visit telehealth consult

Hims defines a subscriber as a customer with at least one subscription that bills automatically, so each new patient who stays is worth recurring monthly revenue. We found no public lifetime value per telehealth patient and will not invent one. Companies selling software to hospitals rather than patients face a different buyer, covered in our guide to healthcare software and AI search.

Where do AI assistants sit in a patient’s search for virtual care?

Before the visit: people ask an assistant about symptoms and options, then many go looking for a provider.

KFF found that about a third of adults used AI for health information in the past year, most often to look up symptoms or general information, and 65% of users cited wanting quick information as a major reason. Rock Health’s December 2025 survey (opens in a new tab) of 8,000 adults, reported by Fierce Healthcare, also found 32% use, with ChatGPT the most used (23%), then Gemini (15%). Its researchers note people also use AI for “looking for specific providers and clinics,” and 40% consulted a provider after a chatbot answer.

The handoff from question to care is where a telehealth brand can be named. The platforms treat that handoff differently:

  • OpenAI says ChatGPT Health (opens in a new tab) “is not intended for diagnosis or treatment” and is designed to support, not replace, care from clinicians. Its launch post does not describe booking visits with outside telehealth providers.
  • Amazon documents the opposite design: its Health AI agent can connect users directly to licensed One Medical providers. Fierce Healthcare reports it reached US consumers through Amazon.com and the Amazon app in March 2026.

Our inference: when the assistant belongs to a company that also sells virtual care, independent brands are competing with the assistant’s owner. When it does not, the assistant answers from what it can find on the web, so a brand’s public evidence matters more. Health apps face a similar test, covered in our guide to digital health apps.

What do patients ask AI before choosing a virtual care service?

Whether a service covers their state and insurance, what it costs, whether it is legitimate and how it handles data.

The prompts below are illustrative, written by us to show the shape of provider-discovery questions. They are not captured from real users, and they deliberately stay away from asking what condition someone has or which treatment to take.

StageIllustrative prompt
Access“Online doctor available tonight in Ohio for a common illness”
Insurance“Virtual therapy services that accept my insurance plan”
Price“How much does a telehealth visit cost without insurance?”
Comparison“Compare subscription telehealth services on price and cancellation”
Legitimacy“Is this telehealth company legit? Are the doctors licensed?”
Privacy“Does this online therapy company share my data with advertisers?”
After the visit“Can I get my prescription sent to my local pharmacy?”

The insurance question is getting louder. Teladoc said demand for insurance-covered services at BetterHelp was “stronger than anticipated” and outpaced available provider capacity in the second quarter of 2026. Our inference: a service that clearly states which plans it accepts, in which states, gives an assistant a precise answer to repeat.

How does an AI answer become a first visit?

Through a short list, a legitimacy check, an intake form and a consult, then a subscription or follow-up care.

  1. Named for a need. The assistant lists a few services for a type of care, a state or an insurance plan.
  2. Checked. The patient asks whether the company is legitimate, licensed and private. In our brand reputation study, 88.0% of AI answers to “Is this brand legit?” cited a review or complaint platform, and 99.7% of complete answers raised at least one problem with the brand.
  3. Intake. The patient fills in an online questionnaire, often on a phone, and pays or confirms coverage.
  4. Consult. A licensed clinician decides what care is appropriate. That decision belongs to the clinician, not the brand or the assistant.
  5. Retained. Subscription models bill monthly, therapy continues over weeks, and pay-per-visit patients return when they need care again.

Step two is where telehealth brands are most exposed. Trustpilot and the BBB made up 61.7% of review-platform citations in that study, so complaints about billing or cancellation can shape the description a patient sees before the brand’s own pages do.

What decides which telehealth brand an assistant names?

Verifiable facts about access, price, licensing and reputation; the platforms do not publish their selection rules.

What is documented: OpenAI and Amazon describe the purpose of their health assistants, as above. Google documents a hard gate for paid search: telemedicine providers must be LegitScript-verified to advertise. No platform documents an equivalent rule for which services its AI answers name.

What has been observed:

What we infer: the trust factors in this industry are concrete and checkable. They include state licensure of clinicians, a clear list of states served, accepted insurance, transparent self-pay prices, LegitScript certification where it applies, a plain privacy policy, easy cancellation and a clean complaint record.

What does it cost a telehealth brand to be missing or misdescribed?

More paid acquisition in a category where marketing already consumes a large share of revenue.

We have no direct measurement of patients lost to AI absence, so we show the economics and label the reasoning. Hims spent $262.2 million on marketing in one quarter. Teladoc cut its advertising and marketing spending from a year earlier, and BetterHelp’s paying users fell. Paid search is restricted to certified providers, and the Federal Trade Commission (opens in a new tab) required BetterHelp to pay $7.8 million after finding it shared sensitive data, including that people had been in therapy, with platforms such as Facebook for advertising. The FTC said that targeting helped bring in “tens of thousands of new paying users.”

Our inference: as targeting with health data narrows and paid search stays gated, being named accurately in answers people already ask becomes a more valuable source of patients. A brand described with an outdated price, a state it no longer serves or an old complaint loses those patients to the next name on the list. The steps for correcting those errors are in how to fix wrong brand information in AI answers.

How does generative engine optimization work for a telehealth company?

Generative engine optimization (GEO) makes your service easy for assistants to find, describe accurately and verify through outside evidence.

For a consumer virtual care brand, the work usually covers:

  1. One consistent identity. The same brand name, services, states and prices across your site, both app stores, review platforms and partner listings, so assistants do not merge you with a similarly named service.
  2. Access facts in text. States served, insurance plans accepted, hours, typical wait times and self-pay prices, kept current and written on pages an assistant can read.
  3. Clinician credibility. Named medical leadership, how clinicians are licensed and supervised, and service pages reviewed by clinicians, without promising any outcome or treatment. Wellness brands need the same care with claims; see how wellness brands stay lawful in AI answers.
  4. Certifications and policies. LegitScript status where it applies, a privacy page that says plainly what you never share for advertising, and clear cancellation and refund terms.
  5. Reputation where patients check. Accurate profiles on Trustpilot, the BBB and the app stores, real patient reviews and complaints resolved in public. Fake reviews also distort AI answers and carry legal risk.
  6. Regulatory currency. Telehealth rules change. The DEA’s fourth temporary extension, reported by Telehealth.org (opens in a new tab), allows clinicians to prescribe controlled substances without a prior in-person visit through December 31, 2026. Pages that explain what your service can and cannot do under current rules keep assistants from repeating old ones. Drugmakers face similar rules on what AI repeats, covered in how pharma brands stay accurate in AI answers.
  7. Measurement. Ask a fixed set of illustrative access, insurance, price and legitimacy questions across ChatGPT, Gemini, Perplexity and Google’s AI features, by state where it matters, and track who is named and which sources are cited.

No telehealth company can be promised a place in AI answers, and anyone who offers that promise is overselling. The aim is narrower: when a patient is ready to be seen, an assistant can describe your service correctly.

What remains unknown about AI and telehealth patient acquisition?

How many virtual visits begin with an AI answer, and how assistants choose among telehealth brands, are not yet measured.

  • Follow-up is not booking. KFF and Rock Health show people contact providers after AI answers, not which provider or whether it was virtual.
  • Company figures mix causes. Hims’ growth and BetterHelp’s decline reflect products, prices and insurance shifts, not AI visibility.
  • No telehealth ranking studies. The Jacques audit studied health information citations, and our reputation study covered brands in general. Applying them to telehealth brands is our inference.
  • Platform designs differ and change. Amazon routes to its own care; OpenAI says its health product is not for diagnosis or treatment. Either could change.

Where should a telehealth company start?

Start with the access, price and legitimacy questions patients in your states ask, and see who assistants name.

That first check usually shows whether your service is named for the types of care and states you serve, whether your prices, insurance and licensing are described correctly, which review sites shape your reputation and which competitors appear instead. From there, the work is to publish accurate access facts, strengthen your reputation and keep every page current with the rules.

If your growth depends on completed intakes and first visits, ask us to review how AI assistants describe your service. We will map where your brand appears for access, insurance and legitimacy questions, why competitors are named instead, and which changes are most likely to bring more patients to your intake. How those changes are carried out, from state, insurance and price facts in text to clinician credibility and reputation work, is set out on our generative engine optimization service page.

Frequently asked questions

Do patients use ChatGPT to find a telehealth provider?

Some do. Rock Health found people use AI chatbots for “looking for specific providers and clinics,” and KFF found 58% of AI users with physical health questions later followed up with a provider.

Can a telehealth company advertise in AI answers?

We found no documented way to buy placement in AI health answers. On Google Ads, telemedicine providers must be verified by LegitScript before they can advertise.

Does Amazon’s Health AI recommend other telehealth companies?

Amazon documents that its Health AI connects users to its own One Medical providers. It does not describe recommending outside telehealth services.

Will health claims help a telehealth brand get named?

They add legal risk without documented benefit. Describe access, licensing, prices and policies accurately, and leave diagnosis and treatment decisions to clinicians.

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