Building BidScribe: Why I'm Tackling RFP Responses with AI

Responding to RFPs is painful. Every B2B company knows it — you get a 50-page document full of questions, and your team scrambles to piece together answers from previous proposals, internal docs, and tribal knowledge. Under deadline pressure.

I'm building BidScribe to fix that.

The Idea

The core insight is simple: most RFP answers aren't novel. They're variations of things your company has already written. The challenge is finding the right previous answer and adapting it to the current context.

This is a perfect use case for RAG (Retrieval-Augmented Generation). Build a knowledge base from your best answers, retrieve the most relevant ones for each new question, and let an LLM synthesize a tailored response.

Where I Am Right Now

Let me be honest: BidScribe is an idea and an early prototype. There's no live product, no users, no landing page. I started working on this over the past weekend, and I'm at the "validating the concept and planning the architecture" stage.

My strategy is to build in public — share the journey, document what I'm learning, build an audience — and launch when the product is actually ready. That's not today.

Planned Tech Stack

I'm going with Next.js + Supabase — a stack I know well and that lets me move fast as a solo developer.

Why Supabase over Firebase? Three reasons:

1. pgvector — vector similarity search natively in PostgreSQL
2. Row Level Security — multi-tenant isolation without application logic
3. Edge Functions — serverless compute without leaving the ecosystem

What I've Learned So Far

Even in the planning and early prototyping phase, I've learned a lot:

The RAG problem is harder than it looks. Building a demo that retrieves relevant text is easy. Building something that reliably handles structured business documents with nuanced queries is a different challenge entirely. I'm writing more about this in my RAG architecture post.

The market is real. Every B2B company I've talked to about this immediately gets it. RFP response is a pain point that people actively spend money trying to solve. That's a good sign.

Solo building is viable — with AI help. I'm using an AI agent setup that significantly accelerates my development. Without it, building this as a side project alongside a full-time job and family would be much harder.

What's Next

The immediate plan:

1. Validate the core RAG pipeline — can I build retrieval that's good enough for real RFP questions?
2. Build a minimal UI — upload documents, ask questions, get draft answers
3. Test with real users — find a few companies willing to try it on actual RFPs
4. Iterate based on feedback — then figure out pricing, packaging, and launch

I'll keep writing about the journey as it unfolds. If you're interested in the RFP response problem or building AI products as a side project, follow along.