You are building an app that will process scanned expense claims and extract and label the following data: - Merchant information - Time of transaction - Date of transaction - Taxes paid - Total…

The prebuilt Read model extracts text lines and words without understanding the document structure or labeling specific fields like merchant or total, requiring significant post-processing to achieve the goal.

A custom template model requires manual labeling of documents to define the fields to extract, which involves more development effort than using a prebuilt model.

A custom neural model requires a larger dataset of labeled documents for training, which involves significant development effort compared to using a prebuilt model for a common document type.

The prebuilt receipt model in Azure AI Document Intelligence is specifically designed to extract common fields from receipts, such as merchant name, transaction date, total, and taxes, requiring no training and thus minimizing development effort for this specific use case.