A telehealth brand gets cited by AI assistants when its site and its outside profiles agree on who the company is, which conditions it treats, who the prescribing clinicians are, which states it serves, and what it charges. The work has four parts: an llms.txt file, schema for the organization, its condition and treatment pages, and its clinicians, plus matching licensure, review, and pharmacy disclosures.
AEO, GEO, AI visibility, and AI SEO are four names for the same work: getting a brand named and cited inside AI answers. This post uses AEO. The [plastic surgery version of this setup](/blog/plastic-surgery-llms-txt-schema-entity-setup) centers on one surgeon and a local practice. A telehealth brand differs in four ways: it sells nationally with no local pack, it serves different states under different licenses, patients ask legitimacy questions before they ask about price, and several brands in a category share near-identical names.
What does the technical layer of telehealth AEO include?
The technical layer has four artifacts: an llms.txt file at the site root, JSON-LD schema for the organization, conditions, treatments, and clinicians, sameAs links tying the brand to outside profiles, and a set of public disclosures covering licensure, states served, and pharmacy partners. Each artifact removes a guess an AI system would otherwise make about the brand.
Content decides whether a page is worth citing. The technical layer decides whether the correct entity gets the credit and whether the brand reads as legitimate. Which sources AI engines pull from is covered in the ClinicAds post on [where each AI engine gets telehealth recommendations](/blog/chatgpt-perplexity-gemini-ai-overviews-telehealth-recommendations), and this post does not repeat that split.
- llms.txt: a plain-language map of the brand, conditions, clinicians, and states at the domain root
- Schema: JSON-LD naming the organization, each condition and treatment page, each clinician, and visible FAQs
- sameAs: links from each entity to the same entity on outside profiles
- Disclosures: licensure, states served, pharmacy partners, and pricing stated in crawlable text
Does llms.txt help a telehealth brand get cited?
Evidence says it probably does not, yet. SE Ranking analyzed about 300,000 domains, found roughly 10 percent had an llms.txt file, and reported no measurable link between having one and how often a domain was cited by major AI models. The finding is correlational, and no major AI provider has committed to reading the file.
ClinicAds still recommends writing one as a clarity exercise and not as a ranking lever. A brand that cannot describe its conditions, clinicians, states, and pricing in one page has an entity problem, and the same sentences feed the schema and the directory text. Jeremy Howard proposed the format in September 2024.
What should a telehealth llms.txt contain?
A telehealth llms.txt should contain the legal business name, a one-sentence definition, the conditions treated with links, the states served, the clinician group with license types, the pharmacy arrangement, the pricing model, and a disambiguation line. Keep the file under 80 lines, written as facts, with no promotional adjectives and nothing the brand cannot document.
The disambiguation line matters because DTC brands in one category often share naming patterns, and an AI answer can merge their reviews or attribute one brand's complaints to another. State the legal name, the trading name, and the domain in one sentence so the file separates the brand from lookalikes.
- Line 1: the business name as an H1, matching the domain, app store listing, and corporate record
- Line 2: a one-sentence definition, such as an online clinic providing prescription treatment for named conditions in named states
- Section: conditions and treatments, each linking to its page
- Section: states served, stated as a list and updated when licensure changes
- Section: clinical team, with each prescriber's full name and license type
- Section: pharmacy arrangement, naming whether prescriptions are filled by a partner or a compounding pharmacy
- Section: pricing model, stating whether pricing is a subscription, a per-order charge, or both
- Section: disambiguation naming the legal entity and the domain
Which schema types does a telehealth site need?
A telehealth site needs four schema.org types in JSON-LD: MedicalOrganization for the company, MedicalCondition or a Service for each condition page, Physician for each clinician, and FAQPage for pages with visible questions. Schema.org defines each type, and the Schema Markup Validator at validator.schema.org checks them. Each block belongs on the page that is about that entity.
MedicalOrganization is described by schema.org as a medical organization that can be physical or not, which fits a company that operates online. It carries properties such as medicalSpecialty and isAcceptingNewPatients, and its subtypes include MedicalClinic and Physician. ClinicAds suggests MedicalOrganization for the company and Physician for each prescriber.
| Schema type | Page it belongs on | Key properties | What it settles |
|---|---|---|---|
| MedicalOrganization | Home, about, contact | name, url, legalName, telephone, medicalSpecialty, isAcceptingNewPatients, sameAs | Which company this is and what it treats |
| MedicalCondition or Service | Each condition or treatment page | name, serviceType, provider, areaServed, offers | Which condition the page covers and who provides care |
| Physician | Each clinician biography | name, jobTitle, medicalSpecialty, worksFor, sameAs | Who prescribes, with what license, for which company |
| FAQPage | Pages with visible Q&A | mainEntity of Question and Answer pairs | Which questions the page answers, in the exact answer text |
How should condition and treatment pages be structured?
Each condition page should cover one condition, state who is eligible, name the treatment options, list the states served, show the price or pricing model, link to the prescribing clinicians, and carry safety information. One page per condition lets an AI system quote a self-contained passage and lets schema describe exactly one thing.
Pricing is where DTC brands can gain ground. A patient asking an AI assistant what a monthly prescription program costs gets ranges and a caution. A brand that publishes its monthly price and what it includes on a crawlable page gives the model a figure to quote with attribution. A brand that shows price only after intake gives the model nothing, and a comparison site fills the gap.
- One page per condition, never one page listing every condition
- Eligibility and contraindications stated in visible text, reviewed by a licensed clinician
- States served listed on the page, matching llms.txt
- Price or pricing model shown in text, with Offer markup only for prices visible on the page
- Clinician reviewer named, with a link to the biography page
How do you make each clinician a findable entity?
Each clinician becomes a findable entity through a biography page carrying Physician or Person schema: full name, jobTitle, license type, medicalSpecialty, worksFor linking to the company's @id, and sameAs links to the state license lookup and professional profiles. Patients ask AI assistants whether a telehealth company uses real, licensed clinicians, and a verifiable page answers that.
Telehealth adds a wrinkle: a clinician may hold licenses in several states and treat patients only where licensed. List the license states on the biography page and keep them consistent with the states-served list. When a clinician leaves, the biography, the schema, the llms.txt team list, and the outside profiles must change together, or an AI answer keeps naming someone who no longer works there.
- name: exactly as on the state license record
- jobTitle: the license type spelled out, such as Nurse Practitioner or Physician
- medicalSpecialty: the specialty the clinician practices
- worksFor: the company, by @id
- sameAs: license lookup page and any professional profile with a stable URL
- Update trigger: a departure or new license changes the page, schema, llms.txt, and listings in one pass
Which licensure and legitimacy disclosures should be public?
The public disclosures should cover the states served, the license type of each prescriber, how prescriptions are filled, and any third-party certification the brand holds. Legitimacy is a common patient question for online prescribing, and a brand that states these facts in crawlable text gives AI systems something verifiable to repeat.
Certification rules vary by market. Google's Healthcare and Medicines policy sets country-specific requirements for telemedicine and online pharmacy advertisers, and LegitScript certification is a common route to approval in several countries. A brand should check the current policy for each market it advertises in and publish a certification only if it holds one. The certification question is also a paid-media gate, covered in the ClinicAds post on [advertising GLP-1 telehealth without getting flagged](/blog/advertise-glp1-weight-loss-telehealth-without-getting-flagged).
Claims on condition pages need the same discipline. Outcome statements should be substantiated and reviewed by counsel before publication, because schema and llms.txt repeat whatever the page says.
Which listings need to match the website?
The listings that need to match are the Google Business Profile if the brand has a physical address, app store listings, Trustpilot and similar review profiles, Reddit and comparison-site descriptions the brand can influence, and pharmacy partner pages. Name, legal entity, states served, treatments, and prices should read the same across them.
The audit is mechanical. List every profile, copy the name, entity, states, treatments, and price into one sheet, and mark every difference. Fix the brand's own pages first, then the profiles it controls, then request edits where it does not. Common conflicts are a retired product name, an old price after a plan change, and a states list that predates a license expansion.
| Field | Site | App store | Review profiles | Comparison sites |
|---|---|---|---|---|
| Brand and legal name | Match | Check | Check | Check |
| States served | Match | Often absent | Check | Check |
| Conditions treated | Match | Check | Rarely listed | Check |
| Monthly price and inclusions | Match | Often absent | Check | Check |
| Clinician names and licenses | Match | Rarely listed | Rarely listed | Rarely listed |
| Pharmacy arrangement | Match | Often absent | Check | Check |
How do you check that the setup works?
A brand checks the setup in three steps: validate every schema block, confirm llms.txt loads at the root with a 200 status, and run a fixed prompt set in ChatGPT, Perplexity, and Gemini each month, logging whether the brand is named and described accurately. Validation proves the markup is well formed. Only the prompt log shows whether the answers changed.
Expect a lag. The month-by-month view is in the ClinicAds post on [how long AEO takes for a telehealth brand](/blog/how-long-does-aeo-take-to-work-for-a-telehealth-brand). Telehealth economics are measured in CAC and payback. ClinicAds plans paid telehealth acquisition at a CAC of $90 to $250 per paying patient, patient LTV of $600 to $2K, payback under 3 months, and a 3:1 or better LTV to CAC ratio on $10K to $50K a month, all as agency averages, not guarantees. Each paying patient an AI answer produces without ad spend lowers blended CAC.
This setup is one part of the search visibility system ClinicAds builds for [telehealth brands](/telehealth), alongside paid media and speed-to-lead. It is not sold as a separate service.
- Validate every JSON-LD block at validator.schema.org and fix errors before launch
- Fetch /llms.txt in a browser and confirm a 200 response and plain-text rendering
- Ask three AI assistants for the brand by name, its lead condition with the words online provider, and a brand comparison
- Re-run the same prompts every 30 days and log changes against each schema and listing edit
Does a telehealth brand need llms.txt to appear in ChatGPT?
No. SE Ranking found no measurable link between llms.txt and AI citations across about 300,000 domains, and brands are named without one. ClinicAds still suggests the file because writing it forces clear statements of conditions, clinicians, states, and prices.
Which schema type should a telehealth company use?
Schema.org defines MedicalOrganization for medical organizations that may be physical or not, which fits an online clinic. Add Physician markup for each prescriber and repeat the same wording on the page so the markup and visible text agree.
Should a telehealth brand publish its prices for AI answers?
Yes where the brand is comfortable. A monthly price and inclusion list on a crawlable page, with matching Offer markup, gives a model a figure to quote. Prices shown only after intake leave comparison sites to fill the answer.