Skip to main content
Audact
See the offer

Integration · Medatixx

Recall and intake, written back to the practice.

Planned · no date set

Medatixx runs in thousands of German Hausarzt and specialist practices. The planned integration writes appointments, recalls and triage notes back into it, while the governance runs in the call: the disclosure, the professional boundary, and the detection of checksum-validated identifiers (BSN, IBAN, card).

What the agent produces per call

{
  "event": "call.completed",
  "outcome": "appointment_requested",
  "outcome_source": "conversation",
  "confirmed_by": null,
  "receipt_id": "rcpt_01HX7K9M3P4Q5R6S7T8U9V",
  "article_50": { "disclosed": true, "clause": "1", "at_turn": 1 },
  "scope": { "pack": "medical_de", "locale": "de-DE", "escalated_to_human": false },
  "write_back": {
    "system": "medatixx",
    "patient_ref": "sha256:e3b0c44...",
    "starts_at": "2026-09-03T08:20:00+02:00",
    "reason": "Rezeptanforderung"
  }
}

Worked example. Acme GmbH and Nova Industries are fictional; no customer data appears here.

patient_ref is a hash. The identifier that maps it back sits in its own compartment with its own key, which is what makes an erasure request answerable.

What applies to the call

The patient must be told they are speaking to an AI system, at the latest at the first interaction.
Regulation (EU) 2024/1689, Art. 50(1) and 50(5)
Health data may only be processed on one of the narrow grounds.
Regulation (EU) 2016/679 (GDPR), Art. 9
The data subject may request erasure, and the controller must be able to carry it out.
Regulation (EU) 2016/679 (GDPR), Art. 17

What Audact governs in the call

Article 50 disclosure at the top of the call, recorded with its position.
In controlled testing
Professional scope: clinical questions route to a person instead of being answered.
In controlled testing
Checksum-validated identifiers (BSN, IBAN, card) are detected; other patient data is not in the tested scope.
In controlled testing
Write-back of appointments, recalls and triage notes into Medatixx.
Planned

Sector

Monday morning is the whole problem.

A Hausarztpraxis takes a week's worth of calls in two hours on a Monday, and the people who can answer them are the people who should be with patients. Prescriptions, appointment moves, results: high volume, low judgement, and almost none of it needs a doctor.

What does need a doctor is the caller with a symptom, and that boundary is the product. The agent takes the appointment; the moment a question becomes clinical, the call goes to a person and the receipt records where it turned.

Setup

Four steps, once the integration ships.

  1. Connect Audact to your Medatixx installation with an API token.
  2. Map agent personas to physicians and appointment types.
  3. Configure the professional scope rules and the disclosure wording.
  4. Run in shadow mode, then switch to active.

Questions

Does the agent triage symptoms?
No. It records what the caller says. A transfer to a person, triggered by the agent, is in controlled testing; automatic hand-off on red-flag signals is in development. Triage is a clinical judgement and it stays with someone qualified to make it.
Where does patient data live?
In the practice system, as before. Audact holds a hashed reference and the fields needed to write the appointment back; the mapping key is compartmentalised, and destroying it on erasure is coming soon.

Do not wait for the integration

Every integration will be built on the same six endpoints. They run inside the managed platform; no public API is served on api.audact.ai yet, apart from the published Merkle roots. Where a sector system exposes an API, the shorter route is to call it from your side and let Audact govern the conversation.

See the API

Related