When a family asks ChatGPT for treatment in your city, are you in the answer?
A growing share of families and patients now ask ChatGPT, Claude, Perplexity, Gemini, and Google's AI Overviews for recommendations instead of scrolling results. Those answers name three or four centers. We track whether you are one of them and fix it when you are not.
What changed
For twenty years, being found meant ranking on a list. Now the list is being replaced by an answer. The engines that write the answer read your site differently from Google's crawler: they need clear entities, structured data, direct answers to direct questions, and consistent facts about who you are across the web. Most treatment center, behavioral health and therapy practice websites have none of that.
The mechanics are not a secret. We wrote them up in how to get recommended by ChatGPT and Google AI. This service is that work done every month for your own programs, with a report that says what changed.
Who this is for
Residential treatment centers and detox, IOP and PHP programs, behavioral health organizations running several levels of care, and therapy practices small enough that the owner still takes the intake call. The work is the same in each case, but the questions are not. A center gets asked about by city and level of care. An IOP gets asked about by hours, insurance and the population it serves. A practice gets asked about by modality, panel and who is accepting new clients. The query list is built around what you actually run. If most of what you do sits on the behavioral health side, this pairs with our website work for mental health and therapy practices.
What we track every month
The queries families actually type
“Best rehab in Phoenix,” “detox near Scottsdale,” “IOP for alcohol use in Arizona,” “EMDR therapist in Tempe who takes Aetna,” and twenty more specific to your programs, levels of care and locations.
The engines that write the answers
ChatGPT, Claude, Perplexity, Gemini and Google AI Overviews. They often disagree, so each one is checked on its own and reported side by side.
Whether you are cited, and who is named instead
Whether each engine mentions you, what it says about your programs, levels of care or clinicians and whether that is accurate, and which centers, programs or practices it names when it does not name you.
Month over month change
The same queries on the same engines every month, so a change in the report reflects a change in the answers, not a reworded prompt.
What we fix
Schema an engine can read
MedicalOrganization, the services you actually deliver, and FAQ markup, so an engine gets your facts as structured data instead of guessing at them from prose.
Entity consistency
Your name, address, programs and description matching across your site, directories and profiles. Engines cross-check facts between sources, and disagreement makes you harder to cite.
Citation-worthy content
Pages built around the questions families actually ask, with the answer in the first sentence, so there is something clear and accurate to quote.
llms.txt
A short summary of your center and its key pages, in the proposed llms.txt format. Support across the engines is still uneven, and it costs little to have in place.
Crawler access for AI engines
Checking that robots.txt, your firewall or your host is not quietly blocking the crawlers AI engines use, such as GPTBot, ClaudeBot and PerplexityBot.
The landing page engines quote from
Levels of care, insurance accepted, location and how to start, or for a therapy practice the modalities, the panels and who is accepting new clients, laid out so the facts an engine lifts are clear, current and easy to find.
What it is not
It is not a replacement for your SEO agency. It runs alongside them. It is not ads. It is the organic side of a channel that did not exist two years ago.
How to get it
The monthly report, fixes included, is part of Growth Partner at $2,500 a month, alongside the full-site build, conversion testing and weekly content. Every tier gets the snapshot at launch; Growth is where it becomes a monthly report.
Either way it starts with a baseline, so the first report is an honest picture of where you stand today, not a result.
Book a 30-minute call with Ali
Frequently asked questions
Generative engine optimization: the work of making your center easy for AI engines like ChatGPT and Perplexity to find, trust and cite when they write an answer. SEO aims for a place in a list of links. GEO aims for a mention inside the answer.
Answer engine optimization. The term is older and grew up around featured snippets and voice assistants, and today most people use it interchangeably with GEO. Both mean being the source an engine quotes when it answers a question directly.
SEO earns a position in a list of results. This earns a mention inside an answer, where there is no list and usually only a few names. The technical groundwork overlaps, which is why it runs alongside your SEO agency rather than replacing it.
The first report is a baseline, not a result. Engines refresh on their own schedules, so give it two to three monthly reports before judging, and expect the movement to be uneven from one engine to the next.
No, and anyone who does is guessing. Nobody controls what a model returns. What we can do is fix what makes you hard to cite on your side, and show you every month, engine by engine, whether it is working.
Ranking well helps, but it does not carry over automatically. A center can sit on page one for a query and still be missing from the AI answer to the same question, because nothing on the page was written or structured to be quoted. The first report shows which case you are in.
The same way a center does, with a narrower question behind it. Somebody asks for an intensive outpatient program in a city, for one that takes their insurance, or for one that works with adolescents, and the engine names a few. We run those queries for your programs every month, record whether each engine names you, what it says about your levels of care and hours, and which programs it names instead. Then we fix the structure and content on your side so there is something clear and accurate to quote. Nobody can make an engine cite you, so the honest part of the answer is that we measure it and work on the part we control.
A shorter query list and more variation between the engines. The questions people ask about a practice tend to name a modality, an insurance panel, a city, or who is accepting new clients, so that is what gets tracked, engine by engine, along with what each answer says about your clinicians and who it names when it does not name you. Whether the answers name practices at all, or a directory or a hospital system instead, is one of the things the first report tells you. The fixes are the same work at a smaller scale: schema, facts about the practice that agree with each other across the web, and pages that answer the question in the first sentence.
