AI Data Extraction from PDFs with Dynamic Airtable Prompts

This n8n workflow implements the Dynamic Prompts pattern for AI data extraction, ideal for handling unstructured data like PDFs (e.g., resumes/CVs). It triggers on changes to an 'input' field or extraction columns in Airtable, reading prompts from field descriptions to instruct an AI model on what data to extract. The workflow processes the PDF input, applies the dynamic prompts to extract attributes such as full name, address, and experience, then updates the Airtable row automatically.

Key benefits include reduced maintenance—no need to edit the workflow when extraction needs change; simply update Airtable field descriptions. It's highly reusable across tables or datasets, cutting costs for multiple extraction tasks. The pattern supports unknown or evolving attributes, making it perfect for HR, document processing, or any variable data extraction.

Use cases: HR teams parsing resumes for candidate databases; invoice processing for finance; shipment labels extraction in logistics; or any scenario with semi-structured docs. Includes video demo and example Airtable base for quick setup. Baserow version available for alternatives.