MCP (Model Context Protocol) is the open standard that lets AI assistants like Claude and ChatGPT-based agents use tools. SalesMind AI's MCP integration lets your AI assistant research and configure your outreach setup directly — campaigns, personas, lead lists, playbooks and analytics — instead of you clicking through the app.
What can my AI assistant do through MCP?
| Task | Example prompt to your assistant |
|---|---|
| Audit configuration | "Which of my personas have empty targeting details?" |
| Research setup | "Which campaigns target fintech companies, and with what solution text?" |
| Configure | "Create a persona for Heads of RevOps at 50–200 person SaaS companies" |
| Improve content | "Rewrite my playbook's pricing section using this document" |
| Analyze | "Summarize reply rates by touchpoint for my running campaigns" |
What does MCP deliberately NOT do?
Message-time decisions. The pipeline that writes, checks and sends your messages is deterministic and fully governed — no AI assistant (yours or ours) makes ad-hoc decisions at send time. MCP is for research and configuration; the sending pipeline enforces the same rules for everyone, every time. This split is intentional: discovery is flexible, execution is governed.
How do I connect?
- Get your MCP endpoint and credentials from your workspace administrator or support.
- Add the endpoint to your assistant's MCP configuration (Claude Desktop, Cursor, or any MCP-compatible client).
- Access is scoped to your agent and your permissions — your assistant can only see and change what you can.
Why this matters for AI-first teams
Your documentation and your configuration surface are both machine-readable: your AI assistant can read these docs, inspect your live setup through MCP, and make changes you approve — closing the loop between "knowing what to do" and "doing it".