HackUTD ยท 2025

2nd Place

AutoPM

Multi-agent copilot that takes a product idea from research to live Jira tickets.

AutoPM screenshot

Problem

Product managers lose hours to busywork (user stories, prioritization, OKR alignment, status emails, ticketing) instead of strategy.

What it is

Each project is a Space where eight Gemini-powered LangGraph agents hand off structured work: idea generation and market research with sources, user stories, RICE prioritization, OKR alignment from uploaded PDFs, interactive HTML wireframes, stakeholder emails via Gmail, and Jira tickets.

Highlights

  • 8 cooperating agents passing Zod-validated, type-safe outputs between graph nodes
  • Produces real artifacts end to end: wireframes, sent emails and live Jira Cloud tickets
  • Market research extracts themes and pain points instead of opaque sentiment scores

Stack

  • Next.js
  • React
  • TypeScript
  • Tailwind CSS
  • LangGraph
  • LangChain
  • Gemini 2.0 Flash
  • Zod
  • Tavily
  • MongoDB
  • Auth0
  • Gmail API
  • Jira Cloud