ConvoOps

Inspiration

Teams lose momentum after meetings because decisions and next steps stay as notes instead of becoming real execution. Existing tools like Fireflies.ai, Avoma, and TaskPilot summarize well, but most stop before action, so humans still do the manual work of creating tickets, sending follow-ups, drafting docs, and scheduling meetings.

What it does

ConvoOps turns transcripts into supervised execution. It ingests a meeting transcript, extracts structured intelligence, proposes an action plan, gets human approval, and then uses MCP-powered agents to execute tasks like GitHub issues, follow-up emails, meeting scheduling, and term sheet docs, with full auditability.

How we built it

We built a layered workflow using FastAPI for APIs, LangGraph for orchestration, and MCP for tool execution. The graph runs ingestion, classification, extraction, and planning in parallel, pauses at a human-in-the-loop supervisor, then conditionally executes approved actions through specialized agents (GitHub, Gmail, Calendar, Google Drive) and records outcomes for traceability.

Challenges we ran into

Getting reliable MCP connectivity across different transports and providers was the biggest challenge. We also had to handle one-shot execution constraints, repeated approval flows, optional field validation (like folder IDs), and UX clarity for reviewing many action items without clutter.

Accomplishments that we're proud of

We built a real end-to-end supervised agent workflow, not just a summarizer demo. The product now supports conditional multi-agent execution, clear review controls, incremental action execution, structured results, and enterprise-friendly audit trails tied to meeting-driven actions.

What we learned

Human supervision is not a limitation, it is the trust layer that makes agent automation deployable. We also learned that robust agent systems need strong state management, explicit validation, transport fallbacks, and UX patterns that keep users in control while still reducing manual work.

What's next for ConvoOps

ConvoOps roadmap focuses on continuous innovation. We are committed to expanding our ecosystem with new enterprise integrations and enhancing our AI capabilities for even more predictive action insights. Future developments include customizable automation workflows and deeper analytical tools to extract strategic trends from meeting data. This ensures ConvoOps remains a leader in transforming conversations into verifiable, impactful business outcomes.

Built With

• fastapi
• langgraph
• python
• react
• tailwindcss
• typescript
• vite