A custom Slack bot that assists a coach by monitoring client channels, drafting responses in the coach's tone using GPT, and combining them with pre-written messages and images. The bot posts drafts for human approval before sending and logs all interactions in Airtable or Supabase to prevent repetition.
High-touch coaching programs often run client support through private Slack channels, where clients post updates, questions and wins throughout the week. Clients expect timely, personal responses, and that responsiveness is part of what they pay for. As a coach takes on more clients, answering every channel thoughtfully consumes hours each day, and replies become shorter, slower or more generic, which clients notice.
Handing replies to an assistant or junior coach introduces inconsistency in tone and advice, and a fully automated bot would feel impersonal and risky in a relationship built on trust. Coaches also struggle to remember which tips, worksheets or images they have already shared with each client, so resources get repeated. The coach needs help drafting replies that sound like them, drawing on their materials, while keeping final say on every message.
A Slack Bolt app listens to client channels and detects new client messages that need a coach response.
It gathers context: the client's program stage and goals from Airtable or Supabase, recent channel history and a log of resources already sent.
An LLM drafts a reply in the coach's tone, using examples of their past messages, and suggests a pre-written tip, worksheet or image from Google Drive or Dropbox.
The draft is posted privately to the coach with approve, edit and discard buttons; nothing is sent to the client without the coach's action.
Approved messages are posted to the client channel and logged, so the same resource is not suggested again and the coach can review history.
A high-ticket dating coach uses this Slack bot to manage daily client communication. The AI drafts a supportive message based on the client's current progress and pairs it with a relevant pre-written tip, allowing the coach to approve or edit the response with a single click, dramatically increasing their capacity.
Slack Bolt SDK, OpenAI API, Airtable/Supabase, Python/Node.js, Google Drive/Dropbox API
Clients often share personal details in coaching conversations, and topics such as relationships, health or sexuality can be sensitive information under the Privacy Act 1988 where it applies. Tell clients that AI helps draft responses and that the coach reviews every message, and describe what is sent to the AI provider in your privacy notice. Check where Slack, your database and the AI provider store data. Coaches who are not registered health practitioners should avoid offering therapy or clinical advice, and the bot's drafting rules should reflect that boundary.
This solution requires careful evaluation of cost-benefit ratio. Consider implementation only after thorough analysis and risk assessment.
No. Every draft goes to the coach privately for approval or editing, and only approved messages are posted to the client channel. That human-in-the-loop design keeps the relationship personal and lets the coach correct anything that is off before the client sees it.
It is given examples of your past messages, your common phrases and guidance on what you would and would not say. Over time, your edits show where drafts miss the mark, and those examples can be added to improve future drafts. It will not be perfect, which is why approval stays in place.
The same approach can be built for other platforms with APIs, such as Discord or some community tools, though approval flows differ. Slack's interactive buttons make the approve and edit step particularly smooth. Setup typically takes one to two months including testing with a few clients.
Get expert guidance on implementing Slack AI Coaching Assistant for your business
Contact Our AI Experts