Skip to content

Chapter 12: Manual and automated evaluations.

Image 12 - Measure

Overview

You can thoroughly assess the performance of your generative AI application by applying it to a substantial dataset. Evaluate the application in your development environment with the Azure AI Evaluation SDK.

When you provide either a test dataset or a target, your generative AI application outputs are quantitatively measured with both mathematical-based metrics and AI-assisted quality and safety evaluators. Built-in or custom evaluators can provide you with comprehensive insights into the application's capabilities and limitations.

Manual and Automated Evaluation Approaches

Effective evaluation combines human judgment with automated measurement:

  • Manual Evaluation: Human graders assess outputs for nuanced quality factors
  • Automated Metrics: Mathematical measures (F1, BLEU) and AI-assisted evaluators (groundedness, relevance)
  • Iterative Testing: Start manual, scale to automated once patterns emerge
  • Domain Expertise: Involve subject matter experts for specialized applications

Azure AI Evaluation SDK

The Azure AI Evaluation SDK provides comprehensive evaluation capabilities:

  • Built-in Evaluators: Groundedness, relevance, coherence, fluency, safety metrics
  • Custom Evaluators: Define domain-specific quality criteria
  • Batch Evaluation: Test against large datasets locally or in the cloud
  • Integration: Works seamlessly with Azure AI Foundry and GitHub workflows

Manual evaluation builds intuition; automated evaluation provides scale. Together, they give you confidence that your AI system performs reliably before it reaches users.

Online Resources

Next Steps

Continue your learning journey:

← Chapter 11 | Chapter 13 →


Questions or feedback? Join the discussion on our GitHub repository or connect with the community.