Built an AI-powered FF&E product data extraction workflow that converts a single product-page URL into structured, ready-to-use product data in Google Sheets. The workflow validates the URL, fetches the supplier webpage, extracts relevant HTML content and the product image, uses Gemini to convert unstructured page content into structured fields, and updates the correct Google Sheet row automatically. It currently captures: SKU, finish/colour, supplier, vendor description, notes, product image, product URL, automation ID, and processing status. I also implemented failure handling and retry logic so blocked, inaccessible, or temporarily unavailable product pages do not leave the workflow hanging. Failed requests are routed separately and the corresponding Google Sheet row is updated with an appropriate status. Tech stack: n8n, Webhooks, HTTP Request, HTML Extraction, Gemini, Google Sheets. This project gave me hands-on experience with AI-assisted information extraction, structured outputs, row-level record matching, retry mechanisms, and production-oriented error handling. I’d value feedback from others building similar n8n, AI extraction, or data automation workflows—especially on how you would further improve reliability across websites with different page structure