Connect folk to your favorite AI tools (Claude, Cursor, ChatGPT…) to read and write your CRM in plain language — no app switching, no manual digging.

Have your agent write a follow-up using your folk context plus your last call, then save it back to folk so the record stays current.

"Draft a follow-up email to the CEO of Notion using our last call, then log it as an interaction."

Your agent retrieves the contact's profile, recent notes, and last interaction from folk (or pulls the recap from a connected meeting tool), drafts a personalized email, and — on your go — logs the sent follow-up as an interaction and adds any next steps as a note.

search_people → get_person → list_past_interactions → create_interaction → create_note

If a note-taker like Grain, tl;dv, or Fireflies is connected directly to folk, those call notes already live on the contact record — so get_person and search_notes give the agent the meeting context without a separate meeting-notes MCP.

folk supplies the relationship history; your meeting tool (or folk's synced notes) supplies the call.

Connect folk to your AI client via the folk MCP, include the contact's name and company, then review, send, and let the agent log it.