Telehealth Therapy Notes in 2026: How AI Documentation Tools Are Built for Virtual Practice

Quick answer (for featured snippet)

AI documentation tools for telehealth work by accepting de-identified session text or transcripts and generating structured SOAP or DAP notes in seconds. For telehealth specifically, notes must also document the platform used, verbal telehealth consent, clinician and client locations, and any audio quality issues that affected observation. PsyFiGPT is designed for exactly this workflow.

This post is part of our complete guide to HIPAA-Compliant AI for Behavioral Health Practices.

Telehealth therapy has stabilized as a permanent part of behavioral health practice. By 2026, most behavioral health clinicians see a meaningful portion of their caseload remotely—and the documentation burden that comes with it has not gone away.

In some ways, telehealth has made documentation harder. Audio quality is inconsistent. The nonverbal cues that inform clinical observation are narrower on a screen. Consent for telehealth must be documented separately. And payers increasingly audit telehealth claims for compliance with platform and location requirements.

AI documentation tools are now mature enough to address most of these problems directly. This guide covers how to use AI for telehealth therapy notes effectively, what telehealth-specific documentation requirements look like, and how PsyFiGPT is built to support virtual practice workflows.

How telehealth documentation differs from in-person notes

The clinical content of a good therapy note—Subjective, Objective, Assessment, Plan—does not change based on session modality. What changes is the set of additional elements telehealth requires.

Telehealth-specific documentation requirements

Most payers and state licensing boards expect telehealth notes to include:

• Platform used — the name of the HIPAA-compliant video platform (e.g., Zoom for Healthcare, Doxy.me)
• Verbal telehealth consent — documentation that the client consented to receiving services via telehealth at the start of the session
• Clinician location — the state and setting where the clinician was providing service
• Client location — the state and physical setting where the client was during the session (required for cross-state licensure audits and some payers)
• Audio/video quality — a brief notation if technical issues affected the session or clinical observation
• Session medium — video only, audio only, or hybrid (relevant if the video connection dropped mid-session)

Failing to document these elements creates gaps that flag during payer audits. Cross-state telehealth has added new complexity: a clinician licensed in one state seeing a client who has relocated to another state must document both locations and confirm they hold appropriate licensure or are operating under an applicable compact.

What disappears in telehealth that documentation must compensate for

In-person clinical observation includes a range of nonverbal data: posture, gait, eye contact, the way a client enters the room. On video, that observation window narrows substantially. Clinicians often rely more on verbal content and what they can observe from the waist up.

Good telehealth documentation compensates by being more precise about what was observed, rather than noting an absence. "Client appeared calm and engaged throughout the session" communicates more than silence on the topic.

Common telehealth documentation mistakes

1. Omitting telehealth consent language

Many clinicians obtain telehealth consent once at the start of care and do not document it in individual session notes. Some payers require per-session consent notation or a reference to a standing consent form. Know your payers' requirements.

2. Using in-person place of service codes

Telehealth claims require specific place of service codes and modifiers. Place of service 02 (telehealth provided other than in patient's home) or 10 (telehealth provided in patient's home), combined with modifier 95 or GT, signals to the payer that the session was remote. Missing these codes triggers claim adjustments or denials.

3. Leaving audio quality issues undocumented

A session where audio dropped for several minutes affects clinical observation. Documenting it—"audio connection was interrupted for approximately 4 minutes mid-session; clinician was unable to observe client's response to [topic] during this interval"—protects the clinician and provides an accurate clinical record.

4. Copy-forwarding from in-person templates without adjustment

Using an in-person SOAP template without adapting it for telehealth often produces notes that are technically complete but missing telehealth-specific fields. AI documentation tools with telehealth-aware prompts avoid this problem.

How AI documentation tools handle telehealth notes

AI documentation tools work with the same core logic whether the session was in-person or remote: the clinician provides session content (in text, via transcript, or via structured recall), the AI generates a structured draft, and the clinician reviews, corrects, and signs.

For telehealth, the workflow has a few additional considerations.

Input: what to provide to the AI

The most common input types for telehealth documentation:

• Post-session recall — clinician writes a brief narrative after the session; AI expands into structured SOAP/DAP format
• Platform transcript — export the auto-generated transcript from your video platform (Zoom, Teams, Doxy.me), de-identify it, and input it
• Mixed input — transcript for Subjective content plus clinician-authored notes for Objective and Assessment

Always de-identify transcripts before inputting. Remove or replace: client name, date of birth, session date (use "session date" or "this session"), unique identifiers, and any location details that could re-identify a client.

Prompting AI for telehealth-aware SOAP notes

A basic SOAP prompt does not automatically include telehealth elements. Modify your prompt to request them.

Telehealth SOAP prompt template:

"You are a clinical documentation assistant for outpatient behavioral health. Using the de-identified session content below, generate a SOAP note. Include the following telehealth documentation fields at the start of the note under a 'Session Details' header: platform used, verbal telehealth consent documented, clinician location (state), client location (state), session medium (video/audio/hybrid), audio/video quality. Then complete the four SOAP sections: Subjective (2–3 sentences on chief complaint and symptom change), Objective (clinical observations, affect, engagement, standardized scores if applicable), Assessment (diagnostic impression and brief rationale), Plan (interventions, goals with timeframes, and next appointment). Flag any content where audio quality gaps affected your observations."

PsyFiGPT supports saved custom prompts, so you can store this telehealth template and apply it with one click rather than typing it every session.

Prompting for DAP notes in telehealth

DAP format is common in psychotherapy for its focus on session content and clinical formulation.

Telehealth DAP prompt template:

"Generate a DAP psychotherapy note from the de-identified session content. Add a Session Details header including: telehealth platform, verbal consent documented, clinician state, client state, and audio/video quality notation. In Data, include client-reported content and clinician observations available via video. In Assessment, formulate the connection between session content and the client's functioning, noting any limits on observation from the telehealth medium. In Plan, list specific interventions discussed, one measurable goal with a timeframe, and any referrals or follow-up items."

Handling audio quality in documentation

Audio quality is one of the most common practical problems in telehealth documentation. Transcripts from poor-quality sessions may be incomplete or inaccurate.

Workflow for low-quality audio sessions

1. Document the issue in Objective — "Audio connection dropped intermittently during the second half of the session; clinician observation of client's verbal responses during this interval was limited."
2. Use post-session recall rather than transcript — if the transcript is unreliable, write a brief narrative from memory and use that as AI input instead.
3. Flag uncertain sections — your AI prompt should include "flag details you cannot verify from the input." Review those flags before signing.
4. Never fill in clinical content you did not observe — do not use AI to construct Assessment or Plan language that goes beyond what actually occurred in the session.

The clinician's signature on a note is a clinical and legal attestation. AI drafts what the clinician instructs and reviews. The judgment remains human.

Cross-state telehealth documentation requirements

Telehealth has created a wave of cross-state licensing complexity. A client who lives in one state and moved during treatment, or who accesses services from a vacation location, can create multi-state licensure questions.

What to document for cross-state sessions

• Client's physical state at the time of service (ask at session start if location has changed)
• Clinician's physical state at the time of service
• Applicable licensure or compact coverage (e.g., PSYPACT for psychologists, LCSW Compact for clinical social workers)

Keep this information in the session record, not just the intake. A client who was in New York at intake may be in Florida for six months of treatment. Your notes should reflect each session accurately.

Intake and scheduling for telehealth practices

Documentation is only one part of the telehealth workflow. The front-end—getting clients through intake, matched to a therapist, and confirmed for their first appointment—carries its own overhead that grows with a telehealth caseload.

PsyFi Assist handles the administrative layer of telehealth practice: intake questionnaire delivery, therapist matching, telehealth appointment scheduling, and session reminders with video links included. For practices that are managing both the documentation and the intake/scheduling burden, combining PsyFiGPT with PsyFi Assist addresses both sides of the operational stack.

See AI Intake and Session Notes: What AI Captures From the First Conversation for more on the intake side, and How AI Intake and Documentation Work Better Together for the integrated workflow.

Telehealth documentation checklist

Use this before signing any telehealth note.

Session Details

• [ ] Platform name documented
• [ ] Verbal telehealth consent noted (or reference to standing consent form)
• [ ] Clinician location (state) documented
• [ ] Client location (state) documented
• [ ] Audio/video quality noted if any issues occurred
• [ ] Session medium noted (video, audio, hybrid)

SOAP/DAP Content

• [ ] Subjective reflects client-reported content from this session specifically
• [ ] Objective observations are limited to what was actually observable via telehealth medium
• [ ] Assessment includes diagnostic impression with rationale
• [ ] Plan includes specific interventions, at least one measurable goal, and next appointment
• [ ] Any audio quality gaps are flagged in Objective

Billing

• [ ] Correct place of service code applied (02 or 10)
• [ ] Correct telehealth modifier applied (95 or GT per payer)
• [ ] Payer-specific telehealth requirements verified for this session type

Using PsyFiGPT for telehealth note workflows

PsyFiGPT is built for behavioral health documentation workflows—SOAP notes, DAP notes, treatment plan updates, and more. For telehealth clinicians specifically:

• Saved prompts — store your telehealth SOAP and DAP templates and reuse them each session without retyping
• PHI de-identification — built-in tools to help identify and remove identifiers before inputting session content
• Encrypted memory — session context stored securely, not shared with model training
• Team plans — group practices can standardize telehealth documentation templates across clinicians

For a detailed look at prompt templates and advanced workflows, see PsyFiGPT Saved Prompts and Team Prompts.

Next steps for your telehealth practice

Telehealth documentation does not have to be slower or more complicated than in-person notes. With a structured AI-assisted workflow, the additional requirements—platform, consent, location, audio quality—take seconds to add, not minutes.

The key is having prompts that already know to ask for them.

Start with PsyFiGPT to try a telehealth documentation workflow today. For the intake and scheduling side of your telehealth practice, see PsyFi Assist.

For HIPAA guidance on AI documentation tools, see Choosing a HIPAA-Conscious AI Documentation Tool for Behavioral Health.