- Active diagnosis with specifier and current severity
- Functional impairment (work, school, relationships, self-care, safety)
- Symptom evidence from this session that supports continued care
- Specific intervention applied (not "supportive therapy" — "CBT cognitive restructuring around catastrophic thinking re: job loss")
What's different about LCSW documentation Clinical social workers carry more documentation weight than most other licensed therapists, for two reasons:
1. Payor mix. LCSWs are disproportionately serving Medicaid, dual-eligible and other public-payor populations where documentation is reviewed closely and audit exposure is real. 2. Format expectations. BIRP and GIRP are the dominant formats — both require explicit intervention and response documentation, which a generic AI scribe will under-serve.
The phrase that should govern your tool choice: the golden thread. Every note must trace from the treatment plan goal → the intervention applied → the client's response → the plan for next session. AI scribes vary widely in how well they preserve this chain — and this is exactly where Twofold leads the category: the Golden Thread is a first-class product feature rather than a byproduct of a good template.
What to look for in a scribe - BIRP native (and GIRP where required), not as a "custom template" you have to engineer. - Golden Thread continuity — the scribe should reference active treatment-plan goals so each note threads to them by default. - Explicit medical-necessity language in the assessment section. "Continued symptoms of MDD with functional impairment in work and social domains" is medical-necessity language; "client doing okay" is not. - Conservative risk language. Over- and under-statement are both audit risks; over-statement is also a clinical risk. - Supervision-friendly output. If you are pre-licensed (LMSW, LCSW-A, ASW), your supervisor reads every note. Narrative-rich notes are easier to supervise from. - Group, family and couples support if you carry those modalities — under-served by most scribes.
Recommended starting points
| Setting | Pick | Why |
|---|---|---|
| Solo LCSW private practice | Twofold | BIRP + DAP + SOAP as defaults; Golden Thread first-class; lowest entry price |
| Cash-pay LCSW, low volume | Twofold or Upheal | Twofold if you want tightest structure; Upheal for a real free tier |
| Established LCSW practice, diverse caseload | Mentalyc | Wider modality library, richer narrative notes |
| Clinic, CCBHC, or SUD program | Eleos Health | Built for behavioral-health orgs; 42 CFR Part 2 attested |
| LCSW-A / LMSW under supervision | Mentalyc | Narrative depth helps supervisors review |
| GIRP-mandated Medicaid work | Eleos or Clinical Notes AI | GIRP native; medical-necessity language tuned for public-payor review |
Medical necessity — what the scribe needs to produce A Medicaid auditor reading a BIRP note expects to see, in roughly this shape:
- Active diagnosis with specifier and current severity
- Functional impairment (work, school, relationships, self-care, safety)
- Symptom evidence from this session that supports continued care
- Specific intervention applied (not "supportive therapy" — "CBT cognitive restructuring around catastrophic thinking re: job loss")
- Client response including any measurable change
- Plan tied to the treatment-plan goal
If your AI scribe drafts notes that consistently miss any of these, configure or switch. Twofold's Golden-Thread-first structure hits the last point (plan tied to goal) by default, which is where more general scribes leak the most.