Exam AI-102 topic 1 question 71 discussion

Explanation: Upload To - Azure Storage Account: Azure Document Intelligence (formerly known as Form Recognizer) works well with images and documents stored in an Azure Storage account, which provides direct integration for processing files. This is the most compatible and streamlined choice for handling scanned images in this scenario. Model Type - Custom Neural: Since the surveys are handwritten and lack a consistent layout, the Custom neural model is the best choice. This model type is designed for flexibility and works effectively with unstructured or semi-structured documents. It can process handwritten text and extract information even when the layout varies across documents, which minimizes the need for custom development and complex template creation.

An Azure Storage account is CORRECT because Azure AI Document Intelligence requires documents to be uploaded to an Azure Storage account to process and analyze them. This ensures seamless integration with the Document Intelligence Studio, enabling you to extract data from scanned images efficiently while minimizing development effort. Custom neural is CORRECT because this model type is designed to handle documents without a consistent layout. It uses machine learning to identify and extract relevant data from diverse and unstructured formats, such as hand-written survey responses. This minimizes development effort as you can train the model on a sample set of your survey responses, and it generalizes well across varying layouts.

from the link - https://learn.microsoft.com/en-us/azure/ai-services/document-intelligence/train/custom-neural?view=doc-intel-4.0.0 Custom Neural supports these document types - Structured surveys, questionnaires Semi-structured invoices, purchase orders