AI-102: How to Streamline Document Processing with Azure AI Composed Models?

Learn how to streamline document processing by combining custom models into a single endpoint using Azure AI Document Intelligence. Simplify workflows and enhance efficiency.

Question

Your organization, Xerigon Inc., processes various types of business documents, such as invoices, receipts, and purchase orders. Each document type follows a different format, and you have trained separate custom models in Azure AI Document Intelligence for each type to extract key information.

You want to streamline the integration process by combining these models into a single endpoint so that the system can automatically determine which model to apply when analyzing a document.

Which is the best approach you should follow?

A. Use the prebuilt document intelligence models to assemble the individual models under a single project.
B. Create a composed model by assembling the individual custom models under a single project.
C. Use Azure Cognitive Search to index all documents and route them to the correct model.
D. Train a single custom model using all document types and their variations.

Answer

B. Create a composed model by assembling the individual custom models under a single project.

Explanation

The best approach in the given scenario is to create a composed model by assembling the individual custom models under a single project. Composed models allow you to combine multiple custom models into a single endpoint. When a document is submitted, the composed model automatically determines which individual model is best suited for extracting data based on the document layout.

To create a composed model using Azure AI Document Intelligence Studio, you would follow the below-outlined steps:

1. Log in to Azure AI Document Intelligence Studio.
2. Click Custom model.
3. Navigate to My Projects, and select one of the custom models.
4. On the left navigation pane, click Models.
5. Under the Models section, select all the models that you would want to include in the composed model, and click Compose (refer to the exhibit).
6. Under the Compose a new model dialog box, provide a Model ID and a Description, and click Compose.

 

Using the prebuilt document intelligence models to assemble the individual models under a single project is not the best approach in the given scenario. Prebuilt models in Azure AI Document Intelligence are designed for specific, common document types such as invoices and receipts. However, they cannot be combined to form a composed model. Composed models only work with custom models created and trained specifically for your organization’s unique document formats.

Using Azure Cognitive Search to index all the documents and route them to the correct model is not the best approach in the given scenario. Azure Cognitive Search is designed for indexing and searching document content and is not optimized for routing documents to specific AI models.

Training a single custom model using all document types and their variations is not the best approach in the given scenario. Combining all document types into a single model compromises accuracy, as the model may struggle to adapt to the unique layouts and structures of each document type. Custom models are better suited for handling specific document formats.

Microsoft Azure AI Engineer Associate AI-102 certification exam practice question and answer (Q&A) dump with detail explanation and reference available free, helpful to pass the Microsoft Azure AI Engineer Associate AI-102 exam and earn Microsoft Azure AI Engineer Associate AI-102 certification.