Clinical message drafter
Draft boundaried, professional replies to client messages and emails, warm, clear, and appropriate to the therapeutic frame.
De-identify first. Remove names and identifying detail before pasting a client message. Use the de-identify tool if you need to.
About this tool
Between-session messages are easy to over- or under-respond to. This tool helps you draft a reply that is warm and contains the message without turning into unpaid therapy-by-text, and that signals the right next step: bring it to session, book a slot, or, if there is any safety concern, the crisis pathway.
De-identify anything you paste, and treat the draft as a starting point. You know the client and the frame; the tool only helps with the wording.
How to use it
- 1Copy the prompt above, or download it as a file.
- 2Open your AI assistant (Claude, ChatGPT, Gemini, or any LLM).
- 3De-identify, then paste the client's message after the prompt.
- 4Choose a version, adjust it to the client and the frame, and send.
The prompt
Copy it, or download it as a file to keep or to save as a reusable prompt in your assistant.
You help me draft a professional reply to a client's message or email. You are a drafting aid for a qualified professional, not a clinician advising on the case. I will paste a de-identified message. Draft a warm, clear, boundaried reply that: acknowledges the client; contains the issue without doing full therapy over text; points to the right next step (bring it to the next session, or book one); and, if the message hints at any risk of harm, gently surfaces the crisis pathway rather than managing risk over text. Keep it brief. Offer a warmer and a more formal version, and remind me to check it against what I know of the client. Here is the message:
What it looks like
[client] emails at 11pm: 'I've had a really bad day and I don't know what to do, can you call me?'
Draft replyWarmer: I'm really glad you reached out, and I'm sorry today has been so hard. I can't offer calls between our sessions, but I want to make sure you're supported tonight. If you feel unsafe at any point, please contact a crisis helpline right away. Let's make your day tomorrow a priority, would an earlier session this week help? We'll pick this up properly then. (A more formal version follows. If there is any hint of risk, the crisis pathway is surfaced rather than handled over text.)
A drafted reply is a starting point, not clinical advice on the case. Check it against what you know of the client before sending.
Frequently asked questions
De-identify it first. Remove the name and any identifying detail before pasting, or run it through the de-identify tool. Never put identifiable client data into a general AI assistant.
If the message hints at risk of harm, the draft gently points to the crisis pathway rather than trying to manage risk over text. You remain responsible for the clinical judgement.
Claude, ChatGPT, Gemini, or any LLM. The prompt works the same in each.
More free tools for psychologists
De-identify client notes
SupervisionSupervision note
ConsentInformed consent form generator
Client materialsPsychoeducation handout
ComplianceDPDP & RCI compliance check
Client materialsCBT thought record
Client materialsDBT skills handout
Professional developmentReflective practice log
Practice operationsBoundary & cancellation messages
A drafting aid for qualified professionals, not a diagnosis, clinical decision, or legal advice. Never paste identifiable client data into a general AI assistant. If you or someone you are with is in crisis, contact a crisis helpline right away.