AI search for physiotherapists: what owners need to know

What patients ask AI when something hurts

Physio questions describe a body part and a hope. Patients ask who is best for a back, who does sports rehab, who can see them today, and whether the clinic is trustworthy. Each question ends with the assistant naming clinics, and the patient books the first that fits:

• “Who is the best and most trusted physiotherapist in the city?”
• “Who should I see for persistent lower back pain?”
• “Are there good physiotherapists near the station?”
• “Which physiotherapists offer sports injury rehabilitation at a reasonable price?”
• “Physical therapist with a same-day appointment for urgent treatment”

Why being named matters for physiotherapists

A physiotherapy patient is a course, not a visit. Someone who trusts a clinic with a back returns for the six or ten sessions it takes, comes back when the shoulder goes, and tells the running club. When an assistant names three clinics to a patient in pain, those three inherit that course and what follows; the other clinics in the area are not in the conversation.

Trust and specificity are what patients ask about, and they are exactly what an assistant can read or cannot. A patient asks who is good with sports injuries or post-surgical rehab, and the assistant answers with clinics whose pages say so. Being one of those names delivers a patient who already believes you are the right fit, which a leaflet in a GP surgery no longer does.

The bad back that books whoever gets named

The decisive moment for a physiotherapist is pain that has lasted long enough to act on. The patient has stopped hoping it will pass and wants someone competent, soon, nearby. They do not want to read clinical bios; they want a name and a reason. An assistant that gives them two or three clinics, each with a specialism and an availability note, ends the search on the spot.

Because the patient is uncomfortable and often anxious, they rarely compare beyond the shortlist. They book the first clinic on it that can see them this week. The clinics an assistant names are on that shortlist. The ones it omits are not, however good their outcomes, and they never learn the patient existed.

How AI decides which physiotherapists to recommend

For physiotherapists, the assistants read the health booking platforms and directories first: in our scans Zocdoc was the most-cited source, Healthgrades second, Practo behind them. They also lean on hospital and university health-system pages, and on the websites of a few clinics that describe their specialisms unusually well. A clinic that is bookable and reviewed on the platforms, with a clear specialism, is a clinic an assistant can name.

Specificity is what separates two well-rated clinics. An assistant asked about lower back pain or sports rehab looks for a clinic whose pages name those conditions and the approach it takes. A clinic that lists "physiotherapy, massage, acupuncture" without saying what it is good at gives the assistant nothing to match, so it names the hospital department or the platform instead. In our scans that happened in two of six cities.

What AI actually says about physiotherapists: our scans

In September 2026 we asked ChatGPT and Gemini six patient questions each, from "best physio in the city" to "same-day appointment for urgent treatment", in London, Dubai, New York, Los Angeles, Chicago and Miami. Seventy-two answers, every clinic and every cited website recorded. The full readout, city by city, is on the physiotherapists directory; the short version is below.

The answers are fragmented and the leaders vary in kind. The assistants named 117 different providers across the six cities. In London an independent clinic group led with 4 of 12 answers; in Dubai a hospital group; in New York a booking platform, not a clinic at all; in Los Angeles a university health system. Chicago's leader was an independent clinic in 3 of 12 answers, and Miami's an independent sports-rehab clinic named in 7 of 12, the highest share of any business in any trade we scan. ChatGPT and Gemini agreed on at least one provider in only two of the six cities, the lowest agreement of the seven trades: a patient who asks a different assistant hears different names.

The sources behind the answers were the health booking platforms first, Zocdoc and Healthgrades well ahead, then Practo, a hospital's own pages and, unusually, two independent clinics' websites, one in London and one in Miami, both of which describe their specialisms in detail. For an owner the reading is direct: your Zocdoc or Healthgrades profile is what the assistants read first, and a clinic website that names the conditions you treat can put you ahead of a hospital department.

Referrals from GPs and coaches do not reach the patient who asks an assistant

Many clinics assume that referrals from GPs, surgeons and coaches are their discovery channel, and that patients who search online are someone else's.

Referrals still bring the patients who go through a doctor first. They do nothing for the patient who wakes up with a back and asks an assistant before anyone else, and that patient is increasingly common. If the assistant cannot read a clear, specific picture of your clinic, it names a hospital group or a booking platform, and the referral network never gets the chance.

What being invisible costs physiotherapists

A missed patient is a missed course of six to ten visits, and often the shoulder and the knee that follow. When an assistant sends them to a hospital department or a platform, you lose the whole sequence, and you never see it: no enquiry, no missed call, just a diary that fills a little more slowly than the year before.

The gap compounds, because named clinics gather more bookings, more reviews and more platform activity, which makes them easier to name again. In our scans one independent clinic in Miami was named in 7 of 12 answers, the highest share of any business in any trade we scan; that is what a legible specialism earns. Staying invisible cedes the patients who search on their own, who are also the ones who pay privately.

For what these odds look like in real cities, the leader's share and how often the assistants agree, see local AI visibility, measured city by city. To see one assistant at a time, the ChatGPT, Gemini and Claude visibility checkers explain what each one reads before it names a physiotherapy clinic.