Skip to main content

AI coaching workspace

An AI assistant for your coaching practice.

Use LaunchSite OS AI to retrieve permitted client context, summarize check-ins, and draft programs, messages, or workflows. Ask for the work you need and review the result with your own coaching judgment.

AI output is assistive. It is not medical advice.
Product previewAI workspace chat with suggested actions such as finding clients who missed a check-in and building an automation

The coaching brain

Every client, fact, and episode,and the links between them.

What the assistant remembers is a map you can open. Each node is a client, a fact you confirmed, or an episode it observed, connected to the others it relates to.

Intelligence map showing client nodes connected to remembered facts and episodes, beside a memory inbox awaiting confirmation

Fictional client data shown for product illustration. No customer information is depicted.

Ask and build

Start from a question, end with the work done.

Which clients missed a check-in this week. Build a training template and assign it. Who is ready to advance a phase. The assistant answers from your data and can carry the next step out for you.

  • 01Answers grounded in your own roster, not a generic model
  • 02Builds templates, automations, and messages on request
  • 03Reference a specific client inline while you type
AI chat home offering to find clients who missed check-ins, build automations, and assign training templates

Your method

Shape how it writes and what it recommends.

Tone, structure, and the boundaries it works inside are yours to set, so drafts come back sounding like your practice rather than like a model.

  • 01Set the voice and structure it drafts in
  • 02Constrain what it may and may not suggest
  • 03Changes apply everywhere the assistant appears
AI customization panel for adjusting assistant tone and behaviour

Where the line sits

It drafts. You decide.

The assistant accelerates writing and synthesis. It does not make clinical or state-changing decisions, and it never presents its own output as your judgment.

  • 01State-changing tools run only from an explicit instruction
  • 02Client-authored content is treated as untrusted input
  • 03Review the facts and source records behind a summary before relying on it
  • 04AI output is labelled as AI output wherever it appears
  • 05Nothing here diagnoses, prescribes, or replaces medical care

Connected AI

Use the AI assistantyou already work with.

A coach can connect Claude, ChatGPT, or another compatible MCP client to LaunchSite OS. MCP is the connection that lets an external assistant discover permitted tools and work with your coaching records. Client support and available tools depend on the connected app and your LaunchSite OS access.

  • 01Connect from your AI app, then sign in to LaunchSite OS and review the authorization request
  • 02Ask for client context, a check-in summary, or supported coaching work using your own records
  • 03Access stays tied to the coach who authorized the connection and the capabilities they can use
  • 04Revoke an individual connection in Settings → Integrations → Connected AI
  • 05The external AI provider receives tool results and handles them under its own privacy and retention terms

Working with AI in your practice

Can AI make coaching decisions for me?

AI can draft, summarize, and carry out supported work you request. You remain responsible for checking the information and deciding how to use it. Generated content can be incomplete or wrong, and it is not a diagnosis or medical advice.

What is the difference between the AI workspace and MCP?

The AI workspace is inside LaunchSite OS. Connected AI gives a compatible external assistant another way to use authorized LaunchSite OS tools. It does not give the external app unrestricted access to your account. Read the privacy policy and data-processing disclosures before connecting sensitive records.

Can my own software connect too?

Yes. The Partner API is a separate integration surface for your own systems and supported workflows. The developer documentation explains API keys, available endpoints, and connection examples.

Your practice, operating clearly

Give it your roster and ask it what you are missing.

Start with a specific client question and check the answer against their record.