AI phone assistant for paediatric practices
Parents call first thing, and a child's age changes how urgent every symptom is. Tel-Agent takes the age and transfers what cannot wait.
- Child's age
- How long
- Fever to a person
Why a paediatric practice loses the morning calls
Paediatric call volume is not spread through the day. It arrives when parents wake to a child who is not right, and it arrives all at once - twenty calls into a twenty-minute window while the waiting room is already filling. An AI phone assistant for a paediatric practice earns its place in those twenty minutes or not at all.
How the AI receptionist handles a call about a child
It answers immediately and asks the child's age first.
It takes what is happening in the parent's words, and since when.
It asks whether the child has been seen recently for the same thing.
If the age and the symptom match a rule the practice set, the call is transferred to a person at once.
Otherwise it offers a slot or takes a callback, and the whole call is written down.
What the agent is told to ask
Written by you, in your own words, as the assistant’s instructions.
What a paediatric phone assistant should and should not decide
What it does
- Answers the morning peak without twenty interruptions
- Asks the age first, because it changes every answer after it
- Records the symptom in the parent's own words
- Applies the practice's own age-and-symptom transfer rules
- Keeps recordings on the practice's own server
What it never does
- Decide whether a child needs to be seen today
- Reassure a parent about a symptom
- Rank one child's urgency against another's
- Give advice about fever, dosage or home care
- Keep anything about a child outside the practice
Questions
Can it decide whether a child needs to be seen today?
No, and it must not try. It collects the age, the symptom in the parent's words and the duration, then applies the practice's own rules about what transfers. Deciding a child is fine to wait is a clinical judgement and belongs to a clinician.
Why is the age asked first?
Because it changes what every other answer means. Practices set their own thresholds - a very young infant with a fever is the usual example - and those are transfer rules, not model decisions.
Parents are stressed. Does an assistant make that worse?
It depends entirely on how fast it hands over. A call answered on the second ring that reaches a person in thirty seconds is better than a phone that rings out during the morning block, which is the realistic alternative.
Can it book an appointment for a sick child?
It proposes a slot and confirms it, unless the age and the symptom match a rule the practice set as urgent - in which case it transfers instead of booking. The rules are the practice's, not the model's.
Does it handle vaccination appointments?
Those are routine bookings and among the easiest calls to take: the child, the age and the appointment. Anything the parent adds that is not routine is flagged for a person.
What about a call from somebody who is not the parent?
Asked and checked. A grandparent or a childminder calling about a child is a common and legitimate call, but what can be discussed differs, so it is routed to a person rather than handled.
Does the EU AI Act allow an AI to answer a call about a feverish child?
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 paediatric practice says on the phone?
You are. A call about a feverish child 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 paediatric 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
