Implement an Azure Document Intelligence in Foundry Tools Solution AI-102 Practice Questions (Page 5)

Questions 21–25

1. 21application · medium

   A law firm wants to train a custom Document Intelligence model to extract case numbers and client names from legal correspondence. They have a folder of 200 scanned letters in PDF format. What is the first step they must take to train the model?

   • ACall the training API with the folder path on the local machine.
   • BConvert the PDFs to a single TIFF file and upload it.
   • CUse the prebuilt general document model to generate labels automatically.
   • DUpload the documents to an Azure Storage blob container and label at least five documents with the fields.
   Reveal Solution Discussion Select an answer first
2. 22expert · hard

   A company processes two very different document types: a one-page application form and a multi-page contract. They have labeled data for both. They want to minimize the number of endpoints they need to manage. What is the best strategy?

   • AUse the prebuilt general document model for both.
   • BTrain one custom model on both document types.
   • CPublish both models and use a separate endpoint for each.
   • DTrain two custom models and compose them into a single composed model.
   Reveal Solution Discussion Select an answer first
4. 23foundation · easy

   You need to extract key-value pairs, tables, and selection marks from a scanned form. Which Azure Document Intelligence model should you use?

   • APrebuilt invoice model
   • BPrebuilt layout model
   • CPrebuilt general document model
   • DPrebuilt read model
   Reveal Solution Discussion Select an answer first
5. 24foundation · easy

   What is the main benefit of creating a composed document intelligence model?

   • AIt improves accuracy for each individual model
   • BIt reduces the number of endpoints you need to manage
   • CIt eliminates the need for labeled data
   • DIt reduces the training time for each model
   Reveal Solution Discussion Select an answer first
6. 25application · easy

   A startup needs to extract key-value pairs from a variety of business cards, but they have no labeled data and need a quick solution. What should they use?

   • AA custom extraction model trained on unlabeled business cards.
   • BThe prebuilt business card model.
   • CThe layout model and parse the text manually.
   • DA composed model of several custom models.
   Reveal Solution Discussion Select an answer first