AI for private clinics: bookings, enquiries and admin in 2026

Most clinic enquiries are admin questions about price, availability and location, and AI can answer those from the clinic's own approved content. Anything about symptoms, suitability or results goes to a clinician.

The safe design is a narrow one. The assistant answers only from a list the clinic has approved, such as the price list, opening hours, parking, which treatments each clinician offers and the cancellation policy. When a patient asks a clinical question, it says a clinician will reply and passes the message to the team. When the patient wants to book, it offers real free slots from the practice system and writes the booking straight into the diary.

That boundary matters for regulation as well as safety. The MHRA's guidance on stand-alone software says software "used to book an appointment, request a prescription or have a virtual consultation is also unlikely to be considered a medical device if it only has an administrative function". Software whose purpose includes diagnosis or treatment can be a medical device. An assistant that stays administrative stays on the right side of that line.

For phone enquiries, an AI receptionist can do the same job by voice. Entry-level UK AI receptionists cost £30 to £100 a month and advanced ones with booking and integrations £100 to £500 a month, according to Fasthosts' July 2026 guide. The guide to AI receptionist costs in the UK compares the published plans.

Send intake forms before the visit and file the answers to the patient record, so clinicians start the appointment with the history in front of them. AI handles the sending, chasing and filing. The clinician reads the history and takes consent.

Review requests should go to every patient after every visit. The Competition and Markets Authority's fake reviews guidance of 4 April 2025 says cherry-picking can happen "by encouraging just those who are satisfied to leave reviews", and it warns against stopping and starting review invitations at will. A filter that only asks patients who rated the visit highly fails that test.

Replies need a person. The GDC's standards tell dental professionals not to post information or comments about patients on social networking or blogging sites, so a reply to a review must never confirm that someone is a patient or describe their treatment.

AI should never give clinical advice in a booking or enquiry flow. Clinical judgement, records accuracy and patient confidentiality stay with the clinic, and each regulator's rules point the same way.

Clinical advice and the MHRA

The MHRA's guidance says a product's medical purpose is judged by what it is presented as doing, and that "general disclaimers (for example 'this product is not a medical device') are not acceptable if medical claims are made or implied". A chatbot that suggests a diagnosis or a treatment cannot hide behind a footnote. Keep the assistant administrative and route every clinical question to a person.

Records and the CQC

Clinics in England that carry out regulated activities must meet the Health and Social Care Act 2008 (Regulated Activities) Regulations 2014, which the CQC enforces. Regulation 17 requires the provider to "maintain securely an accurate, complete and contemporaneous record in respect of each service user". Every automated message and form should write back to the patient record, so the record shows what was sent and when.

Dentists, doctors and confidentiality

The GDC's Standards for the Dental Team require dental professionals to "protect the confidentiality of patients' information and only use it for the purpose for which it was given". The GDC says that covers personal details, medical history, treatment and how much it costs. For doctors, the GMC's guidance Confidentiality: good practice in handling patient information has applied since 25 April 2017, and the GMC last updated it on 13 December 2024.

Health data under UK GDPR

The ICO lists "data concerning health" as special category data, and it can include details of the health care a person receives. The ICO's own example is that a list of osteopath appointments or physiotherapy invoices would count, because it reveals treatment. The ICO says processing special category data needs an Article 9 condition, so agree the condition, the AI provider's data terms and where the data is stored before anything goes live.

Aesthetics marketing and the ASA

Botox and other botulinum toxin products are prescription-only medicines, and CAP Code rule 12.12 bans advertising them to the public. The ASA's advice says no reference to a prescription-only medicine should appear in a sponsored ad, on a website homepage or in logos, testimonials or hover text. A clinic can promote a consultation instead. Any AI that drafts social posts for an aesthetics clinic needs that rule built in, plus a person's approval before posting.

Yarigai is a done-for-you AI agency in London. It builds lead capture and booking flows, support agents that answer from a client's own documents, inbox triage and document reading and filing inside a client's own accounts, which for a clinic means its practice system, inbox and website. The clinic owns the code, the data and the accounts from day one.

A project starts with a free 30-minute scope call on Google Meet, and the clinic keeps the notes whether it hires Yarigai or not. A blueprint of about a week, credited to the build, sets out a written plan, a clickable prototype, the monthly running cost, a fixed price and the payback maths. The build takes 2 to 6 weeks with a demo of working software every week. A single workflow, such as two-way reminders, can go live in about two weeks.

A person approves anything that sends money, emails a customer or deletes data. AI and hosting bills are paid by the client at cost, and the running cost is on the quote. Yarigai's guarantee is one sentence: "If the build isn't profitable for you, we won't build it." Yarigai has no published case study with a clinic. Its support and booking work is described on the AI customer support and websites and apps pages.