School Bid Intelligence System: Dual Gmail-Triggered PDF Processing → AI Bid Classification → Airtable Database & Notifications

Problem:

A contractor pursuing school district bids was receiving bid documents via email across two separate Gmail accounts, manually reading each PDF, classifying the bid type, extracting key information, and logging opportunities into a tracking system was consuming significant time and causing bids to be missed or filed inconsistently.

Solution:

Built a dual-path n8n bid processing system, Path 1 (primary) is Gmail-triggered, extracts text from the PDF attachment, processes it through two sequential Groq LLM chains for classification and data extraction, routes through a Switch node based on bid type, merges classified outputs, sends a Gmail notification with bid summary, and creates structured records in three separate Airtable bases for different bid categories; Path 2 (secondary Gmail account) is independently triggered, runs through a Groq LLM Chain with structured output parsing, fields the extracted data, routes through its own Switch, and creates records in two additional Airtable bases with Gmail alerts, both paths operating in parallel across different inboxes.

Result:

Delivered a fully automated bid intelligence pipeline that monitors two Gmail accounts simultaneously, processes every incoming bid PDF through AI classification and extraction, and logs structured opportunity records into the correct Airtable database with instant email notifications, ensuring no bid is missed and every opportunity is categorized, extracted, and tracked automatically.