Microsoft Certified: Azure AI Engineer AssociateImplement natural language processing solutionsMedium

A logistics company receives thousands of customer feedback emails daily. They want to automatically route these emails to the correct department (e.g., 'Shipping', 'Billing', 'Returns') and identify emails requiring urgent attention. To achieve this, they need to train a model to categorize emails based on their content and assign a priority level. Which Azure AI Language feature should the company use?

  1. AAzure AI Language - Custom Text Classification
  2. BAzure AI Language - Sentiment Analysis
  3. CAzure AI Language - Named Entity Recognition (NER)
  4. DAzure AI Language - Key Phrase Extraction
Show answer & explanation

Correct answer: A. Azure AI Language - Custom Text Classification

Custom Text Classification allows the company to train a model with their own labeled email data to automatically categorize incoming emails into specific, custom-defined categories like 'Shipping' or 'Billing', and even assign priority levels based on content. This directly addresses the need for automated routing and prioritization.

Why the other options are wrong

  • B. Sentiment Analysis determines emotional tone (positive/negative), not the functional category or priority of an email.
  • C. NER extracts specific entities (e.g., names, dates), but not overall email categories or priority.
  • D. Key Phrase Extraction identifies important concepts, but doesn't classify the entire document into custom categories or assign priority.

Azure AI Language Custom Text Classification

A feature of Azure AI Language that enables users to train custom machine learning models to classify text documents (e.g., emails, articles, reviews) into user-defined categories based on their content.

  • Requires labeled training data (text examples with their corresponding categories).
  • Allows for creation of domain-specific classification models.
  • Useful for content organization, routing, and filtering.

Memory trick: Give it a label, know where it belongs.

More Implement natural language processing solutions questions