Test-Driven Prompting

Test-driven prompting adapts Test-Driven Development (TDD) for AI-assisted workflows. Tests become specifications that guide AI generation and automatically verify correctness.

The Test-First Workflow

1. Write tests first — Define expected behavior
2. Share tests with AI — Tests become the specification
3. Generate implementation — AI writes code to pass tests
4. Run tests — Verify correctness automatically
5. Iterate — Fix failures, add edge cases

Why Tests Make Better Prompts

Tests are precise specifications:

• Unambiguous expected behavior
• Clear input/output examples
• Edge cases explicitly defined
• Success criteria built-in

Compare:

• Prompt: "Write a function to validate passwords"
• Tests: Show exactly what valid and invalid passwords look like

Implementing Test-Driven Prompting

Generate tests first: "Write tests for a password validation function that:

• Requires minimum 8 characters
• Requires one uppercase letter
• Requires one number
• Rejects common passwords"

Then request implementation: "Now write the function that passes all these tests."

Benefits

• Clarity — No ambiguity about requirements
• Verification — Automatic correctness checking
• Edge cases — Explicitly covered in tests
• Refactoring safety — Tests catch regressions

When to Use

• Complex business logic
• Functions with many edge cases
• Code that must be correct (payments, security)
• When natural language specs feel ambiguous