Back to projects

Grove Insights

AI-moderated user interview tool for qualitative user research at survey scale.

Overview

I've always believed the best way to solve product problems is to talk directly to users. But I've also felt the friction of manual user interviews—scheduling sessions, coordinating calendars, and spending hours each week conducting and transcribing. Grove was built to give teams the qualitative insights they need without that headache.

Key Features

  • AI-moderated interviews with customizable agents for each study
  • AI-powered insights that surface key learnings from individual interviews and across the full study
  • Chat interface to explore and dig deeper into interview responses
  • Full transcripts and video recordings for every session
  • Google OAuth for simple sign-in
  • Technical Details

    Built with Next.js and OpenAI's APIs for custom voice agents and AI-powered features. Supabase for the backend.

    Learnings

    Learned how to build voice-first custom agents, implement RAG for context-aware responses, and use evals to systematically improve agent performance over time.