🧠 IntelliHQ — Your AI-Powered Team Command Center

🚀 Inspiration

Startups and fast-paced teams live in tools like Slack, but crucial updates, blockers, and decisions get buried in endless threads. We wanted to build a one-click, AI-powered HQ that summarizes team knowledge and surfaces what matters most — without making anyone dig for it.

💡 What it does

IntelliHQ connects to your Slack workspace, ingests recent messages, and uses GPT-4 to generate concise summaries. It identifies key decisions, blockers, and questions, giving your team a clear snapshot of what's happening — all from a clean web UI.

🛠️ How we built it

We divided the 12-hour build into focused tracks:

Slack Ingestion: Used Slack Web API to fetch messages and store them locally in JSON.
Backend: Built with Node.js to expose /get_summary and /ask_question endpoints.
LLM Integration: Crafted custom prompts and connected to OpenAI’s GPT-4 API to process Slack threads.
Frontend: Developed a simple React interface to input tokens, trigger summarization, and display results.
Polish: Added error handling and light styling with TailwindCSS to prepare for demo day.

🧱 Challenges we ran into

Slack API rate limits and workspace permissions delayed testing.
Prompt engineering was tricky — we had to iterate to get meaningful summaries.
React-Flask integration required some debugging due to CORS and token handling.

🏆 Accomplishments that we're proud of

Built a working MVP in under 12 hours.
Successfully connected Slack → Node.js → GPT-4 → React pipeline.
Got accurate summaries of real Slack threads — with GPT identifying blockers we missed!
Designed a usable UI that non-devs could operate.

📚 What we learned

Prompt design is as important as code when working with LLMs.
Real-world APIs (like Slack) bring non-obvious edge cases.
Clear separation of concerns across frontend, backend, and AI layers helped us iterate fast.