AI visibility
Does ChatGPT recommend my therapy practice? What to check before you guess
A calm way to inspect whether AI assistants mention your practice, which public sources shape the answer, and what the result can—and cannot—tell you.
A prospective client can now ask an AI assistant for a therapist in their city, describe a situation, add an insurance or private-pay preference, and receive a short list of names. That makes one question newly important for private practices: does your practice appear?
The tempting response is to open one chat, type one prompt, and treat the answer as a final score. That feels definitive, but it is not a dependable measurement method. AI answers can vary by product, model, phrasing, location context, account, and date. The useful question is not simply whether one answer mentioned you. It is what pattern appears across repeated, relevant questions—and which public sources the assistants rely on when they explain the answer.
Start with questions a real prospective client might ask
A useful check begins with the situations, services, and practical constraints your practice actually serves. Clinical labels matter, but prospective clients often describe what is happening in ordinary language before they know the modality or diagnosis they need.
The questions should also respect the practice's real scope. If you do not provide weekend appointments, accept a particular plan, or work with a given population, a prompt that assumes otherwise measures the wrong market.
- A couples therapist in your city for partners paying privately
- A trauma therapist offering in-person appointments near a specific neighborhood
- A therapist accepting a particular insurance plan and seeing adolescents
- A group practice with current availability for a defined specialty
Record the answer, the position, and the evidence
A name appearing in an answer is only one signal. Save the full response, note where your practice appears, and inspect any links or citations. An assistant may know the practice mainly through a directory, a local listing, an old clinician profile, or the practice's own website.
That source matters because it shows which public description is currently doing the talking. If an outdated directory profile is the only cited source, the immediate task may be consistency and ownership—not publishing another generic article.
Do not turn absence into a diagnosis
Not being named does not prove that the website is poor, that the practice lacks authority, or that a particular technical fix will cause inclusion. It tells you that this assistant, for this question, at this time, returned a different set of names.
Before making changes, compare that result with your Google visibility, Business Profile, directories, site structure, referral activity, and current capacity. AI visibility is one layer of the public discovery system, not a replacement for the rest of it.
Directional is useful when it is labeled honestly. The goal is to notice a repeatable pattern—not to pretend every client receives the same answer.
Measure the trend, not the emotional reaction
Use a stable set of questions and repeat them on a regular schedule. Note new mentions, disappearing mentions, citation changes, and whether improvements to your public information are followed by movement over several weeks.
A weekly cadence is usually frequent enough to catch change without encouraging daily anxiety-checking. Keep the record so a surprising answer can be compared with prior runs instead of becoming an urgent, unsupported conclusion.
Choose one fix that improves the public picture
The best next action is usually concrete and verifiable: correct an inconsistent specialty, strengthen a service page, align a directory profile, complete a Business Profile field, clarify a location, or connect a useful page through internal links.
After the change is live, record the date and keep monitoring. Search and AI systems do not necessarily reflect website changes immediately, so the work needs a review window rather than a same-day promise.
Earshot runs the same client-style questions across connected AI providers, preserves the answers and citations, and places them beside Google, Business Profile, directory, and website signals. The first scan gives you a baseline before you decide whether to activate weekly monitoring.