Feedback Intelligence Agent
Multi-agent system that helps Product Managers go from raw customer feedback to prioritized roadmap gaps. CrewAI agents handle sentiment scoring, theme extraction, roadmap alignment, and gap detection.
Product Managers are drowning in feedback from support tickets, chat, GitHub issues, and app reviews. Reading through hundreds of items manually does not scale. Even when they do, it is hard to quantify which issues hurt most, which themes keep recurring, and whether the roadmap actually addresses the pain.
This project builds a multi-agent pipeline using CrewAI. Specialized agents handle distinct stages: a sentiment agent that scores pain intensity on a 0-10 scale (because sentiment and pain are not the same thing), a theme agent that clusters feedback into actionable groups, a mapping agent that aligns themes to roadmap items using embedding similarity, and a gap agent that flags high-pain themes with no roadmap coverage.
The output is a prioritized list of gaps with LLM-generated recommendations. The kind of report a PM can take into a planning meeting and use to make decisions about what to build next.
- multi-agent
- crewai