A more helpful first conversation.
Planned voice assistance for routine enquiries, appointment coordination and a clear handoff to your team.
Explore the story
Medrella roadmap. This experience is being planned; scope and readiness will be demonstrated before implementation.
What this workflow connects.
The scope below guides discovery and demonstration. This capability is on the Medrella roadmap.
- 01
Routine enquiries and structured intake
- 02
Patient matching and confirmed booking details
- 03
Actions against an authoritative schedule
- 04
Human handoff and failed-action recovery
Start with the calls people make every day
Patients call to ask about availability, confirm a visit, understand where to go or change an appointment. These conversations are repetitive for staff but important to the person calling.
A useful voice agent should resolve a defined set of routine tasks and recognise when a human needs to take over. Medrella's planned voice front desk is designed around that practical boundary. It is an administrative assistant, not an autonomous clinical triage service. The conversation should make its purpose clear and avoid implying that the caller is speaking to a doctor.
Connect speech to a dependable action
Understanding a sentence is only one part of booking an appointment. The system must identify the intended service, use the authoritative calendar, confirm the patient's details and avoid creating duplicates.
A caller may change their mind or correct a detail midway through the conversation. The action should be confirmed in understandable language before it is committed. If the calendar is unavailable, the agent needs an explicit alternative such as collecting a callback request. It should never invent a booking or claim that a change succeeded when the underlying transaction failed.
Design the human handoff before launch
A voice service needs a route for unclear requests, unsupported languages, sensitive matters and callers who simply prefer a person.
Handoff should carry relevant administrative context so that the caller does not need to begin again. The service must distinguish a queued request from a completed transfer. Opening hours, callback expectations and unavailable staff should be described honestly. Clinical concerns should follow the hospital's approved routing language rather than improvised advice. These boundaries are part of the product experience, not small print added after a demonstration.
Use the conversation responsibly
Recording, transcription and retention need a defined purpose and an agreed consent approach. Access to call content should match staff responsibilities, and patient information should not be collected merely because the model can extract it.
The implementation must establish which provider handles the call, where records are stored and how staff can review or correct extracted details. Schedula is the planned foundation for the new Medrella product; voice workflows and clinical handoffs require their own development and verification. Language support, telephony behaviour and call concurrency must be tested in the intended environment rather than inferred from a model's general abilities.
Judge the complete service experience
A pilot should include interruptions, corrections, background noise, an unavailable slot and an explicit request for human help. Review whether the final calendar state matches what the caller understood.
Measure completed administrative tasks, avoidable transfers and staff correction effort together; a short call is not necessarily a successful call. Use synthetic scenarios before considering real patient conversations. Medrella's voice agent remains a planned capability. The website describes the workflow and evaluation approach, while actual availability, language coverage, usage terms and integration scope must be agreed for the deployment.
Connected thinking.
What would more time
for care make possible?
Let's explore it together