Relevance AI
How to automate Relevance AI
Native integrations
Relevance AI has built-in integrations for CRM, communication, productivity, support and analytics tools, and says agents can be triggered from more than 1,000 integrations. Start with the system your team lives in, such as HubSpot or Slack. An idea: an agent that researches every new inbound lead and writes a summary onto the contact.
API and webhooks
Agents can be triggered by a webhook or by a direct HTTP request, and you can add your own API keys to call any REST endpoint from a tool step. A Relevance API key gives programmatic access through the SDK. An idea: have your own product post an event to a webhook when a trial starts, so an agent can prepare a welcome email for review.
MCP server
Relevance AI has an official MCP server at mcp.relevanceai.com. It lets Claude, ChatGPT, Cursor and other clients work with your agents, tools and knowledge, and there is a Claude Code plugin for building and managing agents. An idea: ask Claude to create a draft agent from a process description, then review it in the Relevance AI editor.
Built-in automation
Triggers are the automation layer. Agents can run on a schedule, from an integration event, from a webhook or from another tool. An idea: a scheduled agent that checks open deals every Monday and posts the ones with no activity in Slack.