Microsoft Certified: Azure AI Engineer AssociateImplement knowledge mining solutionsHard
A legal firm is building a knowledge mining solution to analyze contracts. These contracts often contain clauses that define specific terms, such as 'The Effective Date shall mean...', 'The Parties hereto agree that...', or 'Force Majeure Event'. The firm needs to extract these specific clause definitions and their associated explanations as structured data. Which Azure AI Document Intelligence model type is best suited for this highly specific and complex extraction task?
- APrebuilt Invoice model
- BPrebuilt Layout model
- CPrebuilt Read model
- DCustom Extraction model
Show answer & explanationAnswer & explanation
Correct answer: D. Custom Extraction model
Custom Extraction models (formerly Custom Document models) are ideal for extracting highly specific and complex data from documents with varying layouts, such as legal contracts. You train these models by labeling your own documents, allowing them to learn the unique structure and patterns of clause definitions.
Why the other options are wrong
- A. The Prebuilt Invoice model extracts data from invoices, not legal clauses.
- B. The Prebuilt Layout model extracts general text, tables, and selection marks, but lacks the semantic understanding required for specific clause definitions.
- C. The Prebuilt Read model performs general OCR, it does not understand clause structures.
Document Intelligence Custom Extraction Model
An Azure AI Document Intelligence model that you train with your own labeled documents to extract custom, structured data fields from documents with varying or complex layouts, such as contracts or specialized forms.
- Requires human labeling of example documents.
- Ideal for extracting data from documents with unique and complex structures.
- Can extract key-value pairs, selection marks, and tables.
- Supports both template and neural (document) models depending on layout consistency.
Memory trick: Custom extraction extracts custom clauses.