Feb 22, 2026

Black Cell BRD Gen

gdg project tech brd gen ai

 

The Problem

Business analysts and product teams waste hours every sprint hunting across Gmail, Slack, and WhatsApp to gather scattered requirements before they can even begin writing a business requirements document. By the time the BRD is drafted, context is lost, decisions are missing, and stakeholders are misaligned.

The Solution

BRD Generation Agent is an agentic AI platform that automatically ingests emails and communications from multiple sources, extracts business requirements using Google Gemini, and generates a complete, structured BRD in seconds—in any format the team needs.

How It Works

The system follows a three-stage pipeline. First, it connects to Gmail via OAuth2 and fetches relevant emails using flexible filters — by sender, subject keyword, date range, or label. Second, a Gemini-powered AI agent reads all fetched emails and intelligently extracts stakeholders, functional requirements, business objectives, constraints, and open questions. Third, it renders the output into a professional BRD in the user's chosen format—Markdown, HTML, JSON, or plain text—ready to share immediately.

Tech Stack

Built on a DOE (Directive → Observation → Experiment) architecture for reliability and repeatability. Gmail ingestion uses the official Google API with OAuth2. AI processing runs on Gemini 1.5 Flash. The backend is FastAPI orchestrating deterministic Python execution scripts. The frontend is a Streamlit UI with multi-filter email selection, template choices, and one-click export.

What Makes It Different

Unlike generic document tools, this system traces every BRD line back to the exact source email. It supports multiple BRD templates, including standard, agile, and custom prompt-driven formats. The DOE architecture means every step is independently testable and debuggable critical for a live hackathon demo.

Impact

Reduces BRD creation time from 2–4 hours of manual work to under 60 seconds. Eliminates cross-platform searching entirely. Produces consistent, structured output every time regardless of how messy the source emails are.

 

This build was uploaded as a hackathon project

Hackathon

HackFest 2.0

View All Projects
Give a star to encourage!Discussion
Start a new conversation!
Login to join the discussion

More Builds by Utkarsh Gaur

aiconeten gdgnoida hackathon safeai buildwithgoogle
#gdgnoida #hackskill #dveloper #hacktoskill #matrix