AI Progress Notes for CBT, DBT & EMDR Therapists (2026)

AI documentation tools are increasingly being used across all modalities of therapy. But what works well for a general outpatient session does not automatically translate to CBT, DBT, EMDR, or other structured approaches — because the note requirements are different, and the clinically relevant content that needs to be captured is different.

This article covers what to look for in an AI scribe if you practice CBT, DBT, or EMDR specifically.

Note: The modality-specific features discussed below require a tool that understands your documentation structure, not just a generic transcription or summarization service.

---

CBT (Cognitive Behavioural Therapy)

CBT sessions have a recognizable structure: agenda-setting, mood check-in, review of previous homework, session work, assignment of new homework, and a brief session summary. A CBT note that does not capture this structure is missing the clinical logic of the session.

What matters in an AI CBT note:

• The cognitive distortion the therapist targeted and the restructuring technique used (Socratic questioning, thought records, behavioural experiments)
• The client's response to the restructuring — did they move? Did they resist? What was the quality of engagement?
• Homework assigned: the specific task, and the client's stated intention to complete it
• Homework reviewed from the previous session: was it done? What did the client notice?

Where generic AI tools fall short: Most summarizers produce a narrative of what was discussed. A CBT note needs to document what intervention was applied and what happened as a result — because that is what determines whether the treatment is working.

---

DBT (Dialectical Behaviour Therapy)

DBT is a more structured modality than most, often delivered in a multi-component format (individual therapy, skills group, phone coaching, consultation team). Individual DBT sessions follow a diary card review, chain analysis for target behaviours, and skills coaching or acquisition.

What matters in an AI DBT note:

• Diary card review findings — which skills were used during the week, what urges or target behaviours occurred, ratings on emotional states
• Chain analysis: the precipitating event, vulnerability factors, links in the chain, and where the client's skills use succeeded or failed
• The skills module being addressed (TIPP, PLEASE, Wise Mind, DEAR MAN, etc.) and the specific component worked on
• Commitment elicited from the client for the coming week

Where generic AI tools fall short: Chain analysis is a structured clinical technique with a specific format. A summarizer that produces "client discussed a difficult week and identified coping strategies" is clinically useless for a DBT case conceptualization. The chain needs to be in the note.

---

EMDR (Eye Movement Desensitisation and Reprocessing)

EMDR has eight phases, and the note requirements vary depending on which phase you are in. Phase 3 (assessment) and Phases 4–7 (desensitisation and reprocessing) produce very different documentation than Phases 1–2 (history-taking and preparation) or Phase 8 (re-evaluation).

What matters in an AI EMDR note:

• Current phase clearly identified
• Target memory or touchstone if in active processing
• SUDS (Subjective Units of Distress Scale) at start and end of session, with direction of movement
• VOC (Validity of Cognition) score for the positive cognition, if assessed
• Body location of disturbance
• Sets of BLS (bilateral stimulation) used: type, speed, duration — and what the client processed between sets
• Installation status if the session moved to Phase 5
• Closure technique used if the session was incomplete
• Any abreactions or significant trauma responses, and how they were managed

Where generic AI tools fall short: An EMDR session can look like an unstructured conversation from the outside. The clinical logic is in the sequence, the SUDS trajectory, and the processing content — which a generic summarizer will flatten into a vague narrative. EMDR notes without SUDS/VOC scores and phase documentation are inadequate for standard of care.

---

What to look for in an AI tool if you practice structured modalities

Does it offer modality-specific templates? A tool that only produces generic SOAP notes will not serve CBT, DBT, or EMDR documentation well. Look for tools that have note structures aligned to how you actually practice.

Can you customize what gets captured? The ability to specify what fields the AI prioritizes — homework review, SUDS scores, chain analysis — determines whether the note is clinically useful or just a word count.

Does it understand clinical vocabulary? SUDS, VOC, BLS, chain analysis, Wise Mind, thought records — a tool trained on clinical language produces better clinical notes than one trained on general speech.

Does it require human review? You are responsible for the clinical record. No AI-generated EMDR note should skip a SUDS score because the tool missed it. The review step is where you catch those gaps.

---

How MyKaya handles modality-specific documentation

MyKaya includes 20+ note formats, including templates structured for CBT, DBT, EMDR, ACT, psychodynamic work, and others. The system is trained on clinical language and understands the structure of structured modality sessions. Standardized screeners (PHQ-9, GAD-7, C-SSRS) live in a separate client profile section alongside session notes, keeping your documentation organized across a long course of treatment.

Try MyKaya free for 14 days.