Exam AI-102 topic 14 question 1 discussion

guys so what is the correct answer

Contoso have a distinct standard for each office. Is the customized form recognizer more appropriate?

"extract text from the financial documents" BC since Azure Form Recognizer fits the use case

BC 100%

B, E https://learn.microsoft.com/en-us/rest/api/aiservices/custom-models/analyze-document?view=rest-aiservices-v2.1&tabs=HTTP https://learn.microsoft.com/en-us/rest/api/computervision/read/read?view=rest-computervision-v3.1&tabs=HTTP

Analysis of Endpoint C Relevance: Pros: Ideal for extracting information from receipt images, which is one part of the requirements. Cons: Limited to receipts and may not be flexible enough for extracting tables and text from a variety of standardized financial documents. Analysis of Endpoint E Relevance: Pros: Flexible and can handle a wide range of document types including PDFs and JPEGs, fulfilling the requirement for processing standardized financial documents that contain fewer than 20 pages. Cons: Does not specifically target receipt information but can still extract text from receipts. Conclusion Answer is E

B, C: Need to extract receipts and tables. So the endpoints format are actually from a 2.x version of the doc intelligence form recognizer API which can be found here: https://learn.microsoft.com/en-us/rest/api/aiservices/custom-models/analyze-document?view=rest-aiservices-v2.1&tabs=HTTP To extract tables, a call to a custom model could be used if the layout one is not between the options.

Selected Answer: BC

BC is the correct answer.

Is this question still available on May 21, 2024?

Is this question still available on May 21, 2024?

B and C are correct as they are part of Azure Form Recognizer.

Please note that these endpoints are part of Azure’s Form Recognizer service, which is designed for extracting text, key/value pairs, and tables from documents. It’s a great fit for your document processing requirements. The other endpoints listed (Option A, D, and E) are part of the Computer Vision service and are more suited for different tasks such as analyzing results of Read operation, describing an image, and running Read operation respectively. They might not be the best fit for your specific document processing needs.

To meet Contoso’s document processing requirements for standardized financial documents, consider the following two API endpoints: /formrecognizer/v2.0/custom/models/{modelId}/analyze: This endpoint is ideal for extracting information from financial documents. It allows you to analyze custom models specifically designed for invoices and receipts. With this API, you can extract both tables and text from documents that adhere to distinct standards for each office1. /formrecognizer/v2.0/prebuilt/receipt/analyze: Designed for receipt analysis, this endpoint simplifies the extraction of relevant data from receipt images. It’s well-suited for handling financial paperwork in PDF or JPEG formats, especially when dealing with documents containing fewer than 20 pages1. By utilizing these endpoints, you can efficiently process financial documents, extract essential information, and enhance your document processing workflow.

B for customization C for prebuilt receipt with minimum efforts

According to Windows Copilot: B) /formrecognizer/v2.0/custom/models/{modelId}/analyze: This endpoint is part of the Form Recognizer service. It allows you to create custom models for extracting structured data from documents. By training a custom model with examples of financial documents, you can extract relevant information such as tables and text. Since the solution needs to process standardized financial documents with distinct standards for each office, using a custom model tailored to your specific requirements is a suitable choice. C) /formrecognizer/v2.0/prebuilt/receipt/analyze: This endpoint is also part of the Form Recognizer service. It is specifically designed for extracting information from receipt images. Given that the solution must handle receipt data, this prebuilt receipt analysis endpoint is a good fit.

Accroding to the problem, we need: The document processing solution must be able to extract tables and text from the financial documents. The document processing solution must be able to extract information from receipt images. So the answer I will choose BC