RFP Intelligence Agent: Automated Search, Validation, Contractor Q&A & Pinecone-Backed QnA Session

Problem:

A contracting business was manually searching for RFP documents, reading through PDFs to assess qualification, sending details to contractors, handling their questions, and maintaining no structured knowledge base of RFP content, a labor-intensive process that slowed down bid response time and missed opportunities.

Solution:

Built a four-flow n8n RFP intelligence system, Flow 1 searches Google Drive for RFPs, downloads PDFs, extracts content, validates fit using an OpenAI Message Model, and routes qualified RFPs to Google Drive while sending an approval email; Flow 2 is webhook-triggered and sends full RFP details to the contractor as a formatted Markdown email; Flow 3 is Gmail-triggered and extracts contractor answers from email replies, chunks them with a Recursive Character Text Splitter, generates OpenAI embeddings, and uploads the content to a Pinecone Vector Store for permanent knowledge retention; Flow 4 provides a live QnA chat interface, using a Question and Answer Chain backed by a Pinecone Vector Store Retriever and OpenAI embeddings to answer questions about any stored RFP content.

Result:

Delivered a complete RFP lifecycle system, from automated discovery and validation through contractor communication and answer storage to conversational QnA against all accumulated RFP knowledge, replacing a fully manual process with an intelligent end-to-end pipeline.