AI-assisted documentation in psychotherapy
AI-assisted documentation means that a session is recorded, automatically transcribed and a draft for the session documentation is created from it. The transcript is not the documentation but its raw material: what is professionally relevant is decided by the psychotherapist. AI shortens the path from the session to a reviewable text — it replaces neither the professional assessment nor the responsibility for the result.
Why the topic matters
The work after the session is not the exception but the rule: organising notes, keeping track of progress, finding documents again. This effort arises when attention is already on the next appointment. This is exactly where automatic groundwork comes in — not to replace the documentation, but to start with a draft instead of an empty field.
How the process works
Recording
The session is recorded. The required prior consent must be in place for this — that is a requirement for the practice, not a setting in the software.
Transcription
The recording is automatically converted into text. The result is a verbatim record of the conversation, not a professional summary.
Session note
A draft of the session documentation is created from the transcript. It is editable and is reviewed, corrected and deliberately approved.
Deletion of the recording
Once transcription has succeeded, the audio file is deleted. The transcript is kept and can be deleted separately from the recording.
What AI can do here
- Convert spoken language into text and separate speakers from each other.
- Create a structured, shorter draft from a long transcript.
- Bring together recurring themes across several documented sessions.
- Organise existing information for a draft application.
What AI does not do here
It cannot decide what is professionally relevant.
Source: FAQ “Who is responsible for AI results?”
It makes no diagnosis and takes no treatment decision.
Source: Product definition on /en/about
It does not replace review — AI output remains a draft.
Source: Features page, section “Sessions & summaries”
A pseudonymous code instead of the real name is not full anonymisation.
Source: FAQ “What is the difference between a pseudonymous code and anonymisation?”
How Psymira implements the process
In Psymira, recording, transcription and session note work together in one process, and the documented session stays accessible in the case context — together with notes, documents, applications and crisis protocols. In day-to-day work, patients are referred to by pseudonymous codes. AI output consists of drafts: you review, edit and take responsibility for the result.
Data protection and professional control
Audio files are deleted immediately after successful transcription; if a transcription does not take place or fails, a hard limit of 30 days applies. Before AI calls, identifying master data are reduced; pseudonymous codes are used instead of real names. Free text such as notes and transcripts may still contain names — control over this lies with the practice. Psymira does not train its own models with practice data.
Full details on sub-processors, processing locations and retention periods on Privacy & security.
Sources
Primary sources for the legal and technical statements on this page. Product statements are marked above with their origin in our own material.
- § 630f BGB (German Civil Code) — documentation of treatment (in German)
Statutory basis of treatment documentation, including the retention period.
- Google Cloud — data governance for generative AI on Vertex AI
Provider information on how customer data is processed on Vertex AI and not used for training.
- Regulation (EU) 2016/679 (GDPR), Art. 9 and Art. 28
Health data as a special category (Art. 9) and processing on behalf (Art. 28).
- BPtK — practice information “Administrative KI in Ihrer Praxis” (February 2026, in German)
Guidance from the German Federal Chamber of Psychotherapists on the responsible use of administrative AI systems — data protection and professional law requirements, EU AI Act, Medical Device Regulation. It explicitly points out that important questions in the AI context have not yet been settled in a legally binding way.