Summary
Tip
See the Text and images tab for more details!
Microsoft AI Code models are specialized for software engineering tasks. An MAI-Code model can help you explore a codebase, plan a change, generate and modify code, diagnose defects, create tests, and iterate on validation results.
In this module, you learned how to:
- Recognize the capabilities and limitations of an MAI-Code model.
- Select an available MAI-Code model in Microsoft Copilot Code or GitHub Copilot.
- Write a coding prompt with a clear objective, relevant context, constraints, acceptance criteria, and validation steps.
- Use the model iteratively to plan, implement, test, and refine a solution.
- Review generated code for correctness, maintainability, security, accessibility, and unintended changes.
An AI coding model is most effective when you combine clear instructions and useful project context with developer oversight. Treat generated code as a proposed implementation and use repeatable checks to establish whether it is ready to use.