AI-900 Practice Question: Describe features of computer vision workloads on Azure

Answer analysis

Option-by-option breakdown

For each option: why learners choose it and why it is or isn't the right answer here.

• ✗

  Only digital PDF receipts with standardized formatting

  Why it's wrong here

  The prebuilt receipt model in Azure AI Document Intelligence is not restricted to standardized digital PDF files; it accepts images such as JPEG, PNG, TIFF, and BMP from cameras or scanners. The model uses optical character recognition (OCR) to read the text and does not require a fixed template, so it can extract merchant name, date, line items, tax, and total from varied receipt layouts, including angled or low-quality photos. Therefore, saying it handles only digital PDFs with standardized formatting is incorrect.

• ✓

  Sales receipts from stores and restaurants, extracting merchant details, items, and totals

  Why this is correct

  The prebuilt receipt model is a domain-specific document extraction tool optimized for point-of-sale receipts from retail stores and restaurants. It identifies and extracts structured fields such as merchant name, transaction date, itemized line items, subtotal, tax, tip (when present), and the final total, returning them as key-value pairs in JSON output. This directly matches the intended use case and the fields described in the option, making it the correct answer.

• ✗

  Medical receipts and prescription records only

  Why it's wrong here

  This option incorrectly narrows the receipt model's scope to medical documents, but the prebuilt model was trained specifically on general sales receipts, not on medical receipts or prescription records. Healthcare paperwork uses different layouts and field semantics, so Azure AI Document Intelligence provides separate prebuilt models, such as the health insurance card model, or requires custom models to extract healthcare-specific data. The receipt model is therefore not limited to medical use and actually targets everyday retail and restaurant transactions.

• ✗

  Electronic bank transfer receipts for financial transactions

  Why it's wrong here

  Electronic bank transfer confirmations are financial records, but they lack the point-of-sale receipt structure that the prebuilt receipt model expects, such as merchant details and itemized purchase lines. The model's field schema is tailored specifically to retail and restaurant receipts, so applying it to bank transfers would yield poor or irrelevant extraction results; those documents require a different AI Document Intelligence model strategy, typically a custom model or another prebuilt model designed for financial statements. Thus, the receipt model does not target bank transfer receipts.

Go deeper

Related to this question