Saved Prompts and Team Prompts in PsyFiGPT: Curated Clinical AI, Practice-Wide

Quick Answer

Quick answer: PsyFiGPT ships with a two-tier prompt library — personal Saved Prompts each clinician curates, and Team Prompts a practice admin maintains for the whole group. Clinicians can fork a Team Prompt to customize without losing the source. Eleven built-in clinical templates ship out of the box, including the May 2026 additions (DAP Note, BIRP Note, Crisis & Safety Plan, Prior Authorization Letter, Insurance Appeal Letter). Audit logs never include prompt content.

Why Generic AI Prompts Fail in Clinical Workflows

General-purpose AI tools yield different outputs depending on phrasing, timing, and model updates. For casual use, that variation is useful. For clinical workflows, it is a liability.

Therapists need consistency. When drafting a Progress Note, you need the output to match your documentation standards, clinical voice, and therapeutic framework. Inconsistent AI output means extra editing work before any note goes into the record.

Serious clinical practices — hospitals and large group practices — do not give clinicians free-form AI access. They provide templates and protocols ensuring consistency across the system. Research from the American Psychological Association shows standardized documentation frameworks reduce errors, improve continuity of care, and simplify audits.

PsyFiGPT implements this principle through a two-tier prompt library: Saved Prompts for individual clinicians, and Team Prompts for practice-wide consistency.

Saved Prompts — How Individual Clinicians Build a Library

Every PsyFiGPT user has personal Saved Prompts — your private prompt recipe box. Create them directly in the sidebar, edit them as your practice evolves, delete what you stop using, and reorder by frequency.

When a prompt produces exactly the clinical documentation you need, save it with one click. The prompt stays in your personal library — no admin visibility, no shared access. This is your curated set.

Common patterns clinicians save:

• Case-conceptualization prompts structure how you articulate presenting problems and clinical formulations for supervision.
• Treatment-modality-specific prompts shape AI output to match your therapeutic framework (CBT, DBT, psychodynamic, etc.).
• Supervisor-debrief prompts organize session observations and clinical questions in the format your supervisor prefers.

You can flag favorite Saved Prompts with a localStorage toggle, surfacing frequently-used ones at the top.

Saved Prompts are entirely yours. If you leave the practice, they move with you. Admins see only metadata (count of Saved Prompts), never the content.

Team Prompts — How Practice Admins Ship a Vetted Set Everyone Uses

Team Prompts are prompts a practice admin creates and maintains centrally, visible to every team member, standardizing how the group approaches specific clinical tasks.

In a group practice, each therapist's Intake Assessment might differ slightly — one leads with mental status, another with psychosocial history; one uses narrative format, another bullet points. All defensible. But inconsistency across the group creates friction: staff adjust to each clinician's style, coverage becomes harder, clients notice the difference.

Team Prompts eliminate drift. An admin sets up a canonical "Intake Assessment" template, and every new clinician uses the same starting point from day one. You are not removing clinical autonomy — therapists always edit AI output before it enters the record. You are standardizing the baseline.

Team Prompts are visible to all team members. Only Team Admin users can create and edit them. Regular team members can use and fork them, but cannot edit the canonical version.

The Fork Affordance — Make It Yours Without Losing Ours

Forking solves a real problem: you find a Team Prompt valuable but need to modify it for your use case. Without forking, you either edit the Team Prompt directly (affecting everyone) or create a new Saved Prompt from scratch (losing the connection to updates).

Forking copies a Team Prompt into your personal library, fully independent. Admin edits to the Team version do not affect your fork; your local edits do not push to the canonical version.

Edge case: if an admin deletes the Team Prompt you forked from, your fork survives with an indicator that the source is gone. Your content is preserved; the link is broken.

Eleven Clinical Templates That Ship Out of the Box

PsyFiGPT ships with 11 built-in clinical templates in the sidebar, ready to use in any chat. These shape AI output to match documentation standards without requiring you to engineer prompts from scratch.

The original six:

• SOAP Note — Subjective / Objective / Assessment / Plan. Foundation of mental health documentation.
• Progress Note — Lightweight format for ongoing therapy sessions; emphasizes intervention and response.
• Intake Assessment — Comprehensive first-visit documentation: mental status, history, presenting problem, risk, impression.
• Treatment Plan — Goals, evidence-based interventions, and measurable objectives. Required by most insurance.
• Discharge Summary — Course of treatment, outcomes, and continuity planning.
• DSM-5 Reference — Diagnostic lookup tool; ask the AI to walk through criteria for a suspected diagnosis.

The five new templates (May 2026):

• DAP Note — Data / Assessment / Plan. Condensed format popular in psychotherapy and training clinics.
• BIRP Note — Behavior / Intervention / Response / Plan. Focuses on observable behavior and clinical response.
• Crisis & Safety Plan — Risk assessment, de-escalation, and safety planning for acute situations.
• Prior Authorization Letter — Payor-facing documentation of medical necessity.
• Insurance Appeal Letter — Clinical rationale for denied services.

All templates are starting points. You review and edit AI output before any note enters your record. Templates shape the output structure only.

Audit and Compliance — What Gets Logged, What Doesn't

Clinical practices need visibility into who managed prompt data. PsyFiGPT maintains detailed audit logs for all Team Prompt and Saved Prompt actions: who created, edited, deleted, or forked, and when.

Critical privacy protection: Audit logs record metadata only. Prompt content is never logged. This is enforced by a dedicated test in the PsyFiGPT codebase (tests/team/test_team_prompts_audit.py). If you save a prompt, the audit records the timestamp, user, and action ("created Saved Prompt") — never the prompt text.

This protects clinicians. In an audit or review, you can demonstrate controlled prompt management without exposing your personal library's substance. Admins verify system security; clinicians retain privacy.

For deeper context on PsyFiGPT compliance, see HIPAA-Compliant AI for Behavioral Health Practices. For prompt tools and transcript fetching, see Integrating Scribe Transcripts into PsyFiGPT. For multi-step workflows, see PsyFiGPT's Tool Framework for Multi-Step Workflows.

A 3-Minute Setup Walkthrough

Step 1: Save your first Saved Prompt. In any chat, click the save icon next to the prompt input, name it, and save. It appears in your Saved Prompts section. Click to load it into new chats; edit, delete, or reorder from the sidebar.

Step 2: Admins set up Team Prompts. Team Admin users navigate to the Team Prompts admin panel (account settings). Create a new prompt by entering title and text. Once published, all team members see it in their sidebar.

Step 3: Fork a Team Prompt. Hover over a Team Prompt in the sidebar and fork it. The prompt copies into your Saved Prompts as an independent copy. Edit freely without affecting the original.

No complex configuration. The sidebar is your interface; your library grows as you and your team curate what works.

PsyFiGPT is HIPAA-compliant clinical documentation software backed by a Business Associate Agreement. For questions about implementation in your practice, contact our onboarding team at hello@psyfigpt.com.