Turns conversations into coordinated action — across humans, systems, and agents.
Pulls decisions, action items, open questions, and commitments from any meeting transcript — via LLM with structured output.
Rule-based engine maps intent types to destinations: task trackers, document stores, or agents that can research and execute.
Every routing decision is logged. Human corrections feed back to improve future routing. The flywheel is built in from day one.
| Intent Type | Example | Default Route |
|---|---|---|
| decision | "We're going with vendor A" | Document store |
| action_item | "Sarah will follow up by Friday" | Task tracker |
| open_question | "Do we need a backup supplier?" | Agent (research) |
| commitment | "I'll send the contract today" | Task tracker |
| info | "Just FYI, the meeting is at 3pm" | No action |
Contributing a connector, discussing the routing protocol, or exploring commercial use — reach out.
hello@lecollectif.ai →