Property Document Reader: Google Drive Watch → Claude Extraction per Document Type → Structured Airtable Records

Problem:

Client-supplied property documents mortgage statements, inspection reports, tax bills, insurance pages, assessments, utility bills sat as PDFs in Drive, and every data point had to be read and keyed into Airtable by hand.

Solution:

Built a Make.com scenario watching a Google Drive folder. Each new file is fetched and routed by document type into one of seven parallel branches, each following the same pattern: Download the file, Sleep to respect rate limits, send it to Claude via an Anthropic API call with a type-specific extraction prompt (mortgage statement, home inspection report, property tax bill, insurance declaration page, home assessment, comparable sales, utility bill), then Parse Claude Output as JSON and update the Properties record. Branches with array outputs go further Iterators fan the parsed arrays into child records, creating Systems, Findings, Assessment History, Visit History, Comparable Sales, and Utility Bill rows and every branch carries its own Retry or Skip error handler.

Result:

Dropping a document into Drive now yields fully structured Airtable data across seven document types within minutes, eliminating manual data entry from the intake process entirely