The AI-Powered BRD Generator was born out of a real pain point at EMB Global: every new enterprise client needed a detailed Business Requirement Document before work could begin, and creating these manually was taking 3–5 hours per client. I proposed and built an AI-powered solution that compressed this to under 30 minutes.
The core UX is a conversational chat interface where the AI (powered by GPT via OpenRouter SDK) acts as a business analyst — it asks structured questions about the project scope, technical requirements, timelines, and constraints, then synthesizes the conversation into a professionally formatted BRD document. I implemented streaming responses so users see the AI typing in real time (using ReadableStream and fetch with streaming), which dramatically improved the perceived performance.
On the document generation side, I built a template engine that maps AI-extracted entities to document sections — project overview, functional requirements, non-functional requirements, tech stack recommendations, and timeline estimates. Clients can customize the template structure and branding before generation.
I handled the full frontend implementation solo: React component architecture, streaming UI, template system, document preview, PDF export, and integration with the OpenRouter SDK. The backend engineer handled the server-side prompt orchestration. The tool was delivered as an MVP in 6 weeks and immediately reduced client onboarding documentation time by 60%.
Responsibilities
Designed and built the entire chat interface with real-time streaming AI responses
Integrated OpenRouter SDK for LLM API calls with custom prompt engineering
Built the document template engine that maps conversation output to structured BRD sections
Implemented PDF export and document preview functionality
Designed the iterative requirement-gathering conversation flow
Sole frontend developer — delivered full flow in 6 weeks end-to-end