Which agents we build
Booking agent: messengers, the website, the phone, and the schedule of doctors and rooms. Confirmation and reminder agent: the day before, in the morning, and a waiting list. Recall agent: follow-up and preventive visit dates that the doctor sets. Document agent: consent forms and contracts from the clinic’s templates, with no medical data. Report to the manager: bookings, no-shows, schedule load and response time, every morning.
What the market shows
A regional 122 service with AI on doctor appointments handled more than 55,760 medical calls in 2024. A clinic chain in Russia put a voice robot on visit confirmations. It cut no-shows, which had been about 2–3% of bookings, and raised the share of calls answered within 20 seconds from 70% to 82% (January 2025). In April 2026 Russia’s health minister named a clinic’s service tasks as a separate area for AI: schedules, transcription of medical notes, chatbots that route requests; our blog has an article about it. In a study of two US health systems, the share of doctors with burnout in one of them fell by 21.2 percentage points in 84 days after AI took over documentation (2025).
How the pilot works
Days 1–2: review of the process and access. Days 3–5: the metric in the contract, the schedule and answers to common questions, agreed with the chief doctor. Days 6–10: the agent works on one process, most often visit confirmations or inbound requests. Days 11–14: measurement, report and a decision on the next step. You need one responsible employee for half an hour a day and access to the schedule.
Limits
The agent makes no diagnoses, does not assess symptoms or scans, and does not recommend drugs: such questions go to a doctor with a flag. The clinic stays the data controller, and we process data on its instructions. Promotion of services is sent by the clinic under its license, with a warning about contraindications.