Why we built it
The phone goes to voicemail at 6pm. The contact form sits in an inbox until Monday. By the time anyone calls back, the patient has already booked somewhere else. Private clinics lose real revenue this way every week, and it almost never shows up as a single dramatic number; it leaks across dozens of small failures, every week of the year.
The receptionist isn't the problem. The problem is structural: one or two reception staff are the single point of contact between every patient touchpoint and the clinic's internal systems. Whatever falls through, falls through expensively. We built MediConcierge so the predictable 80% of inbound enquiries handles itself, and the 20% that needs a human gets a human, faster.
Who's behind it
MediConcierge is built by RSE Labs, a UK software studio that builds custom platforms for problems too specific for off-the-shelf tools. We work with two or three clients at a time. MediConcierge is what came out of looking closely at how UK private clinics actually run, and it gets the same engineering discipline as everything else we ship: real safety constraints, real tenant isolation, no AI-startup theatrics.
Two of our published research papers go into the engineering substance: how we make a clinical-context AI safe, and how the operational mechanism actually returns labour hours to the clinic. They are the two halves of why a clinic adopts this and the place to start if you want the long-form version of what we are doing.
What MediConcierge is, and isn't
MediConcierge is a clinic operations platform. Patient concierge, CRM, scheduling, and insights, all built around the working assumption that a private clinic's commercial mechanism is hours of clinician time well spent. The AI sits inside the platform where it earns its keep, not as a feature in search of a problem.
It is not a replacement for a practice management system. It does not store clinical notes, write prescriptions, or bill patients directly. It does not replace your receptionist; it takes the predictable load off them so they can spend their hours on the work that actually needs a human. We are deliberately strict about that line because crossing it makes the system worse, not better.
The AI patient concierge will refuse anything that looks like a clinical question. It will not diagnose, contradict a clinician, or volunteer a worked answer to "what could this rash mean". It escalates instead, with a concrete next step rather than a dead end. A patient asking what their symptom might be does not get a chatbot guess; they get a same-day clinician callback offered before they had time to type the question into Google.
Where it runs
MediConcierge runs on Vercel for hosting and on Neon Postgres in AWS' eu-west-2 region (London) for the database. Patient data lives in the UK. Each clinic gets its own logical tenant in a shared schema, isolated by row-level security at the database layer in addition to the application layer.
AI inference is via the Anthropic API. Standard Anthropic infrastructure is US-based; the legal mechanism for that transfer is Standard Contractual Clauses and the UK International Data Transfer Agreement. We are honest about that on the security page rather than implying inference happens in the EEA when it does not. If you would like a longer conversation about data residency for your clinic specifically, that is one of the things the demo call is for.
Talking to us
The fastest way to see whether MediConcierge fits your clinic is fifteen minutes on a video call. We set the AI up against your real schedules, run last week's missed enquiries through the system, and you see the bookings it would have caught. No commitment, no card. Book a demo.
If you would rather just email, hello@mediconcierge.ai reaches a human within the working day.
