"Find a dermatologist nearby with an appointment this week."
A practice website should answer the underlying questions clearly: speciality, location, opening hours, insurance accepted and how to book. This helps patients and services that organise information for them. Whether an AI assistant recommends your practice or can book an appointment also depends on its service, permissions and the booking system.
How Patients Find a Practice#
Search engines, platforms such as Doctolib or Jameda, and personal recommendations remain relevant routes. Your own website complements them. Keep its information consistent with your business profile and booking service so that patients encounter current opening hours and working links.
AI assistants add another route. Their future share of medical searches is uncertain. You can focus on foundations that are useful today regardless of that development.
What AI-Ready Means for a Practice#
1. Publicly Readable Practice Information#
Provide clear text about specialities, services, opening hours, location, languages, accessibility and insurance accepted. Opening hours only displayed in an image are harder to read and maintain. Search services and AI assistants can read HTML; a dedicated API is not a prerequisite.
Structured data can supplement these details. Use appropriate Schema.org types and only information visible on the page. Google does not support every type as a special search result. ICD codes do not replace understandable service descriptions and are not required for general discoverability.
2. A Reliable Booking Route#
Link to the booking system you actually use. Test the link on mobile and explain which appointments can be booked online. Public availability must not expose names, diagnoses or details that identify individual patients.
If your provider supports a suitable interface, availability or booking steps may be integrated. Direct agent booking requires tested permissions, identity checks, confirmation and error handling. A machine-readable booking link alone does not provide those capabilities.
3. Questions Before a Visit#
Is a referral required? What should a patient bring? Is there parking? Which languages does the team speak? Clear answers may prevent routine questions. That is a potential benefit, not evidence for a universal 20 to 30 percent reduction in calls. Compare the number and type of enquiries before and after changes if you want to measure the effect.
Privacy: Separate Public Information from Patient Processes#
Health data receive special protection under GDPR Article 9. They do not belong in publicly available website text, structured data, agent descriptions or unprotected interfaces. Removing a name does not automatically anonymise a record.
This is not a general ban on processing patient data in digital or AI-assisted workflows. A specific processing activity needs an appropriate Article 6 legal basis and, for health data, an Article 9 exception. Necessary processing for healthcare may be permitted under the applicable conditions, including professional confidentiality. Medical secrecy and the roles of service providers also need examination.
Public information interfaces and protected booking processes therefore have different jobs. Transfer only necessary data to the intended system when booking. Check access controls, encryption, logs, retention, deletion and processor arrangements where applicable. An EU server or the label "GDPR-compliant API" does not replace that assessment.
If AI is used in patient communications, examine model transfers, training and retention separately. A practice FAQ can limit its answer source to public information. A bot should not diagnose, request unnecessary confidential details or treat medical questions as routine administration. Provide an appropriate contact route; it is not an emergency channel.
Visibility Without Guarantees#
Google does not require a special AI file, special schema or an API for its AI features. Normal requirements for accessible, indexable and helpful pages still apply. Structured data can support understanding and certain displays, but guarantees neither improved ranking nor inclusion in an AI answer.
A maintained website makes practice information easier to use. Measure whether it reduces calls or generates more suitable booking enquiries. Good technology still needs current content.
The Technical Layers#
The foundation is an ordinary website with clear text and working contact routes. Appropriate structured data can supplement it. An API is an additional option when a specific recipient can use it.
WebMCP is a separate approach for exposing website functions as tools to compatible agents. A file named agents.json is not equivalent to WebMCP and is not a universal requirement for AI visibility. Supported clients and functions such as booking handoff must be tested in the actual project.
What StudioMeyer Does#
We first examine existing content, booking routes and privacy boundaries. Then we determine whether structured data, a public information interface or an appropriate agent function is useful. Scope and cost depend on existing systems and available interfaces.
Start by correcting opening hours and services, making booking clear and keeping patient data out of public layers. Extensions should deliver a demonstrable benefit and work with the systems actually used.
Sources and review date#
Reviewed on 30 September 2026.
