Medical, dental, and wellness practices give AI assistants something to work with when their websites answer the questions patients ask before booking: who they will see, what that provider focuses on, how insurance and self-pay work, and how to get an appointment. Those answers have to come from the practice and pass review by the right people. A list of credentials and a menu of procedure names leave both the patient and the assistant guessing.
Picture someone with a sore shoulder typing a long question into ChatGPT late at night. They want to know which practices nearby see this kind of problem, whether they need a referral, and whether to call a physical therapist or an orthopedic office first. Whatever answer they get can draw, at least in part, on what practices in their area have published.
This guide is for practice owners and marketing leads. It covers content and marketing decisions, not clinical guidance, and nothing here is advice about anyone's care. Where a federal rule touches health marketing, we quote the regulator and suggest bringing the details to your compliance lead or counsel, since state rules also vary.
What do patients ask before they book a practice?
The questions are rarely about the procedure alone. A prospective patient wants to know whether this practice sees people like them, who they would actually see, whether the visit is affordable, and how quickly they can get in. Many also want some sign that other patients had a reasonable experience.
Those questions sort into a short list: provider fit, the first visit, cost and coverage, access and booking, and trust. A dental practice, a chiropractic office, and a behavioral health group will answer each one differently, but each one needs an answer. Ask the people who answer your phones a plain question: "What do callers ask in the first minute that our website could have answered?" Their list is often a more useful starting point than a keyword export, because it comes from real patients in your market.
Keep the questions patients actually ask separate from the ones your team imagines they might ask. Both have a place in planning. The real ones deserve attention first, and they also tell you which words patients use, which are rarely the words on the services menu.
Why isn't a credentials list or a procedure menu an answer?
Most practice websites already carry two kinds of content: a list of degrees and memberships for each provider, and a list of services. Both are usually accurate. Neither answers the question a patient is holding when they arrive on the page.
Consider a fictional dental practice in Ohio whose services page lists implants, veneers, clear aligners, and whitening, each with a sentence or two of general description. A patient comparing practices still does not know what a consultation involves, what the dentist needs to evaluate, how cost is discussed, or who they would see. A procedure name is not an explanation. The page names the service without telling anyone how this practice approaches it, and a general description could sit on any dental site in the country.
Provider pages have the same gap. Initials after a name tell a patient about training, but not which patients the provider usually sees, how they explain options, or which languages they speak. Our guide to provider bio pages that tell patients who they will see covers what to gather and how to make credentials checkable. For services, the companion piece on treatment pages that explain care without promising outcomes shows how to describe what is involved while keeping clinical claims within what the practice can support.
Why does health content carry a higher bar?
Google's guidance on helpful content names a category of topics where the stakes are higher: "We call these "Your Money or Your Life" topics, or YMYL for short." Health sits in that category. The same page says that among the qualities its systems look for, "trust is most important," and that its systems give even more weight to content that aligns with strong experience, expertise, authoritativeness, and trustworthiness on topics that could have a major impact on people's health.
Google's Search Quality Rater Guidelines, the instructions it gives human evaluators, go further: "For pages about clear YMYL topics, we have very high Page Quality rating standards." The same document says no single rating directly changes how a page appears in Search. Read both as a description of what Google says it values, not as a ranking formula.
For a practice, the practical consequence is about accuracy and responsibility rather than tactics. Somebody qualified has to check clinical statements before they go live, and somebody has to notice when they go out of date. A marketing writer should not have to invent what the practice's clinicians know; that knowledge has to be gathered from them and checked by them. The guide on who reviews health content before and after publishing lays out how to split that work between marketing, clinicians, operations, and compliance.
Which pages do the most work for a practice?
A practice does not need hundreds of articles to answer patient questions well. It needs a small set of pages that each do one job and are kept current. The table below is a planning view, not a template every practice must follow.
| Page | Patient question it answers | Who should approve it |
|---|---|---|
| Provider bio | Who will I see, and do they see people like me? | The provider, plus whoever manages credentialing |
| Treatment or service page | What is involved, and is it worth asking about for my situation? | The clinician who performs or supervises it |
| Insurance and cost | Can I use my coverage, and what will I pay? | Billing or revenue cycle lead |
| New patient access | How do I get in, and what do I need? | Front desk lead or practice manager |
| Reviews and patient stories | Did other patients have a reasonable experience? | Privacy officer or compliance lead |
Cost is often the question practices avoid, and it is one patients raise early. The post on insurance, payment, and cost content covers network information, self-pay, and the federal good faith estimate notice. Access is the other quiet gap: whether you take new patients, referral needs, forms, language help, accessibility, and telehealth. The guide to new patient access content treats each of those as a fact the practice states, with a date and an owner.
Where do reviews and patient stories fit?
Reviews matter to patients, and in healthcare they come with legal limits most industries do not face. The HHS Office for Civil Rights has announced a settlement with a dental practice that disclosed patient information while replying to online reviews. In that announcement, the OCR Director said: "Providers cannot disclose protected health information of their patients when responding to negative online reviews."
Testimonials raise separate questions. HHS guidance on the HIPAA Privacy Rule says it "requires an individual's written authorization before a use or disclosure" of protected health information for marketing, with limited exceptions. The FTC's rule on consumer reviews and testimonials, at 16 CFR Part 465, addresses fake reviews, buying reviews that express a particular sentiment, and certain kinds of review suppression. Our post on patient reviews and testimonials walks through each one, framed as questions to bring to your privacy officer or counsel.
How can a practice check what AI says about it?
Start with the questions from your front desk list and ask them in the assistants your patients use, phrased the way a patient would phrase them, with your city or neighborhood included. Record the date, the exact question, the location, the answer, and any sources the answer links. One response on one day is a single observation, not a verdict, so repeat the check over time and compare like with like.
Then look at access before rewriting anything. OpenAI's publisher FAQ says: "For your site content to be included in summaries and snippets in ChatGPT, make sure you aren't blocking OAI-SearchBot." Google's guide to its generative AI features says a page "must be indexed and eligible to be shown in Google Search with a snippet," says a site must also be included in Search generative AI features in Search Console, and adds that "You don't need to write in a specific way just for generative AI search." Neither document promises that any page will appear.
Here is a fictional example. A physical therapy clinic in Colorado asks an assistant which clinics nearby treat running injuries and take a particular insurance plan. The answer names two other clinics and links their pages, which describe the injuries they see and list accepted plans with a date. The fictional clinic's own site says "sports medicine" and "most major insurance." One reasonable hypothesis is that this is a content gap: the clinic has the answers, but it has not published them in a form a patient, or an assistant, could use. Our article on why ChatGPT may not recommend a business covers other reasons an answer might leave a business out, including access and accuracy problems.
Where should a practice start?
Pick one service that matters to the practice and one type of patient. Gather the questions that patient asks before booking, then check whether the provider bio, the service page, the cost information, and the access details answer them. Fix the page closest to the booking decision first, and give it an owner and a review date before moving on.
Ownership matters more in healthcare than in most industries because the facts change often. A provider leaves, a plan drops out of network, a professional society updates its guidance, a location changes its hours. A page with a named owner and a review date has a fair chance of staying accurate. A page without one tends to drift until a patient points out the error.
This is how Hudson South approaches the work. The Authority Asset gathers the practice's approved knowledge and evidence: services, providers, credentials, policies, and what the clinicians are prepared to say in public. Hudson Twenty-One records how the practice communicates across 21 dimensions, so the writing sounds like the people patients will meet. Market Intelligence researches the questions patients in the practice's market are asking. AEO Creation turns those inputs into answer-first content. In health content, the practice's own reviewers should still sign off before anything goes live.
The general version of this method, outside healthcare, is in how to get your business found in AI search. The healthcare version adds one rule that does not relax: accuracy and review come before reach.
See where your practice's answers fall short
On a free 30-minute call, Hudson South walks you through the gaps between your website and the practices patients compare you with.
Get your free competitor analysisSources
- Creating helpful, reliable, people-first content. Google Search Central. Read 2026-10-15.
- Search Quality Rater Guidelines (General Guidelines, September 11, 2025). Google. Read 2026-10-15.
- The dental practice responded to reviews on social media by disclosing patient health information in violation of the law; OCR warns others against this practice. U.S. Department of Health and Human Services. Read 2026-10-15.
- Marketing (HIPAA Privacy Rule guidance). U.S. Department of Health and Human Services. Read 2026-10-15.
- 16 CFR Part 465, Rule on the Use of Consumer Reviews and Testimonials. Electronic Code of Federal Regulations. Read 2026-10-15.
- Publishers and Developers FAQ. OpenAI Help Center. Read 2026-10-15.
- Optimizing your website for generative AI features on Google Search. Google Search Central. Read 2026-10-15.