AI-103: Developing AI Apps and Agents on Azure

Because the invoices exist only as scanned images, there is no extractable text layer for the indexer to use directly. The built-in Optical Character Recognition (OCR) skill in Azure AI Search reads printed and handwritten text out of image content during the enrichment process, and that extracted text can then be projected into searchable index fields, enabling full-text search over the invoice data. Text Split only chunks already-extracted text, Text Translation converts existing text between languages, and Image Analysis produces image tags/captions/visual descriptions rather than extracting the actual textual content needed for searching invoice fields.