AI phone assistant for medical practices
Answer every call during consultation hours. Tel-Agent takes the reason, checks the calendar and transfers the urgent ones - on your own server.
- Reason for the call
- Emergency transfer
- Health data stays put
Why a medical practice needs an AI phone assistant
A practice phone is busiest at exactly the times nobody can pick it up - the morning block, the hour after lunch, the ten minutes before closing. A patient calls, hears it ring out, and calls the practice down the road. The usual answers are an answering machine nobody listens to twice, or a call centre that has to be told what your practice does and still cannot see your calendar.
How the AI receptionist handles a call to a doctor's office
It answers on the first or second ring, in the language the caller is speaking.
It asks who is calling, what it is about and how long it has been going on.
If the answer matches a rule the practice marked urgent, the call is transferred to a person immediately rather than handled.
Otherwise it offers an open appointment slot or takes a message with a callback number.
Everything is written into one searchable archive on the practice's own server - who called, when, what they wanted, what was agreed.
What the agent is told to ask
Written by you, in your own words, as the assistant’s instructions.
What a medical phone assistant should and should not decide
What it does
- Answers during consultation hours, at lunch and after closing
- Records the reason for the call in the patient's own words
- Transfers immediately on anything the practice marked urgent
- Offers appointment slots and takes structured messages
- Keeps recordings and transcripts on the practice's own machine
What it never does
- Decide clinically whether a patient needs to be seen
- Triage a symptom or rank one caller above another
- Give advice about medication, dosage or treatment
- Write directly into a live appointment calendar
- Send a recording to any server outside the practice
Dentists
Pain today, recalls next year - the two calls a dental practice cannot afford to miss.
- Which tooth
- Swelling to transfer
- Recall calls
Dermatology
A waiting list measured in months, and the few calls that must skip it.
- Waiting list
- Skin check or acute
- Referral
General practice
Monday morning, when the same three requests arrive two hundred times.
- Sick note
- Repeat prescription
- Monday peak
Paediatrics
The morning rush, where the child's age changes the answer.
- Child's age
- How long
- Fever to a person
Physiotherapy
A prescription is six appointments - book them in one call, not six.
- Prescription number
- How many sessions
- Series booking
Psychotherapy
A months-long waiting list, and callers who should never hear a machine mishandle them.
- Crisis to a person
- Waiting list position
- Self-pay or insured
Radiology
The questions that decide whether the scan can happen at all.
- Which scan
- Metal or implants
- Claustrophobia
Questions
Is a patient's call health data under the GDPR?
The reason for a call to a practice usually is. Article 9 treats data about health as a special category, which needs a stronger legal basis than ordinary personal data and makes every extra processor a bigger decision. That is the difference self-hosting makes here: there is no extra processor.
Does the assistant have to record the call?
No. Recording is a setting. Tel-Agent can keep only the written transcript, or only the structured message - name, reason, callback number - and no audio at all. Whatever it does keep stays on the machine you installed it on.
Can it tell an emergency from an appointment request?
It can be told which words mean transfer immediately and which mean take a message, and routing rules run before the model does. What it must not do is decide clinically. It hands the call to a human; it does not triage.
Does it replace the practice staff?
It answers the calls nobody could get to - during a consultation, at lunch, after closing. The calls a person would have answered are still answered by that person, because the rules run first and a human line takes precedence.
Can it work with our practice software?
The agent reaches other systems over a generic HTTP tool and webhooks, which is how it fits practice software that has an API. Tel-Agent is pre-alpha and none of this is installable yet.
Does the EU AI Act allow an AI to answer an appointment call?
Yes, under Article 50 of Regulation 2024/1689, which asks that the caller be told they are speaking to a machine - that line belongs in the greeting you configure. Annex III names emergency healthcare triage, so anything that ranked callers by how ill they sound would be a high-risk system. The distinction the Act draws is between assisting and deciding, and this agent assists: it never judges how urgent a symptom is. Austria's KI-Servicestelle at RTR publishes the same reading of the transparency duty.
Who is the data controller for what a caller to a medical practice says on the phone?
You are. An appointment call is health data, which Article 9 treats as a special category, and Tel-Agent is self-hosted with your own model keys, so the recording, the transcript and the record it writes stay on your server and never reach Dpro GmbH. There is no processor agreement to sign with us because we never receive the data, and the source is AGPL-3.0, so you can check that rather than take our word for it.
Can a medical practice run this in the United States?
There is no single federal AI statute to comply with, but a phone line carries older rules. A US practice is under HIPAA, and HIPAA is where self-hosting pays: there is no business associate agreement to sign with us, because no call reaches us. Consent to record is set state by state, and California and Illinois among others want every party to agree. Running the software yourself makes you the caller of record, so the disclosure and the recording switch are yours to set.
Not usable yet. The day it answers a real call will be a release, not a newsletter.
Watch the repository
