--- title: "How Partner Clinics Can Be Discovered by AI, Without Becoming a Marketplace Free-for-All" description: "AI discovery should make clinic information more accurate and usable, while keeping clinics in control of what agents can see and book." image: "https://www.vascue.io/images/blog/how-partner-clinics-can-be-discovered-by-ai.png" canonical: "https://www.vascue.io/blog/how-partner-clinics-can-be-discovered-by-ai" --- [All articles](/blog)Healthcare AI # How Partner Clinics Can Be Discovered by AI, Without Becoming a Marketplace Free-for-All Vascue TeamAugust 15, 2026Updated August 21, 20264 min read ![How Partner Clinics Can Be Discovered by AI, Without Becoming a Marketplace Free-for-All](/images/blog/how-partner-clinics-can-be-discovered-by-ai.png) As people begin asking assistants to find care, clinics need a way to be discoverable in structured, accurate terms. Vascue is treating this as a research design. Partner-clinic discovery and agent booking are not publicly live today. That does not require turning healthcare into an uncontrolled marketplace. ## Discovery Is Not Referral Advice A discovery system can answer administrative questions: which participating clinics offer physiotherapy nearby, which have an opening, and which services are enabled for online booking. It should not claim that one clinician is medically better for a particular patient, infer a diagnosis, or sell placement disguised as advice. The distinction matters. A booking system can make information accessible without pretending to make a clinical recommendation. ## Clinics Set the Rules The clinic should decide what appears: - Locations and public contact details. - Services offered for agent booking. - Participating practitioners. - Availability and booking rules. - Whether bookings are automatic or require clinic approval. - Deposits, cancellation terms, and other pre-visit requirements. Vascue should respect those settings as the source of truth. If online booking is disabled in the clinic system, an agent should not invent a workaround. ## Start Coarse, Then Gate Sensitive Details Public discovery can be useful without exposing too much. A search can show a clinic's broad area and services. Rich details such as live appointment times should be available only after a patient enters a verified, authorized flow. This approach reduces scraping, limits misuse, and prevents an open directory from becoming a source of sensitive operational data. ## What a Discovery Record Actually Contains It helps to be concrete about how little a public discovery record needs. For a participating clinic, coarse discovery can be answered from information the clinic already publishes on its own website: a clinic name, the district or suburb it serves, the broad service lines it offers (physiotherapy, for example, rather than a full price list), and whether agent booking is currently enabled. What it does not contain is just as important. No practitioner diaries. No live availability. No patient counts, no revenue figures, no configuration details from the practice-management system. Anything that would let a third party reconstruct a clinic's schedule is held behind the verified booking flow described in [from "find me a physio" to a confirmed appointment](/blog/from-find-me-a-physio-to-confirmed-appointment). ## How an Assistant Finds the Record Discovery works on two levels, and it is worth separating them. The first is site-level. Vascue already publishes machine-readable descriptions of itself: [llms.txt](/llms.txt), a `.well-known` discovery file, an MCP server card for public knowledge search, and Markdown versions of pages for agents that request them. Those surfaces describe what Vascue is and let an assistant read the public website cleanly. [Building for browsers, search engines, and agents](/blog/building-for-browsers-search-engines-and-agents) explains each of them. The second is clinic-level, and this is the research part. A booking assistant would call a dedicated find-clinics tool exposed through a controlled gateway, pass a location and a service, and receive only the participating clinics that match. That tool is part of the design described in [what an MCP booking server actually does](/blog/what-an-mcp-booking-server-does); it is not publicly deployed. Reading about Vascue on the open web is available to any assistant today. Finding and booking a specific partner clinic through Vascue is not. ## Clinics That Do Not Opt In Nothing happens to them. A clinic that has not joined is not listed, is not described, and cannot be booked through the gateway, even if an assistant asks for it by name. There is no scraped directory behind the scenes and no "claim your listing" model. Participation is an explicit choice made by the clinic, with its own settings, and it can be reversed. That is a deliberate difference from consumer directories, where a practice often discovers it has a profile only when a patient mentions a wrong phone number. ## Keeping Ranking Honest When several clinics match a request, something has to decide the order. The honest answer for an administrative system is to use administrative criteria: distance to the requested area, whether the requested service is enabled, whether there is availability in the requested window. Clinics are not ranked on clinical quality, because a booking system has no legitimate basis for that judgement, and placement is not for sale. If an assistant's own model adds commentary about which clinic is "best", that is the assistant's behaviour, not a signal Vascue provides. The gateway returns facts the clinic chose to publish and nothing more. ## Control Builds Better Participation Clinics will participate only if they can pause, change, or approve the flow. Patients will use it only if the result is accurate and the booking is clearly theirs to approve. Clinics interested in shaping the design can [register on the research page](/ai-agents) without connecting a system. The [AI Front Desk](/ai-front-desk), which handles enquiries and bookings on the clinic's own channels, is the live product this research sits beside. The objective is not to remove people from care. It is to remove the unnecessary friction between an expressed need and a clinic's real availability. This article is part of the [Agent-ready healthcare (research)](/ai-agents) cluster. Start with the pillar page for the product overview, then come back for the detail. Related reading ## Keep reading - ![](/images/blog/building-for-browsers-search-engines-and-agents.png) Healthcare AI ### Building for Browsers, Search Engines, and Now Agents Why websites increasingly need machine-readable discovery, clean content, and controlled APIs alongside their normal user interface. 4 min read [Learn more](/blog/building-for-browsers-search-engines-and-agents) - ![](/images/blog/from-find-me-a-physio-to-confirmed-appointment.png) Healthcare AI ### From 'Find Me a Physio' to a Confirmed Appointment A safe, patient-led example of how an AI assistant could discover and book a clinic through Vascue. 4 min read [Learn more](/blog/from-find-me-a-physio-to-confirmed-appointment) - ![](/images/blog/the-guardrails-behind-ai-booking-in-healthcare.png) Healthcare AI ### The Guardrails Behind AI Booking in Healthcare AI booking should be an administrative assistant with clear limits, not an autonomous actor with broad access to healthcare systems. 5 min read [Learn more](/blog/the-guardrails-behind-ai-booking-in-healthcare) Part of the [Agent-ready healthcare (research)](/ai-agents) cluster[All articles →](/blog)