Step 3: TDD with AI — Keeping You in the Driver’s Seat

Objective
Use AI to accelerate Test-Driven Development (TDD) without surrendering design intent or engineering judgment.

The goal is not to let AI write your tests blindly. The goal is to use AI as a thinking partner while you remain the architect of the code.

Learning Path

1. Re-establish the TDD Loop

Before introducing AI, anchor on the classic cycle:

1. Red – Write a failing test

2. Green – Write the simplest code to pass

3. Refactor – Improve design safely

AI should support this loop, not bypass it.

Key rule:

Tests define intent. AI assists implementation.

2. Use AI to Generate Test Ideas

AI is excellent at producing test scenarios you may not immediately think of.

Ask AI questions like:

Generate unit test scenarios for this function.
Include edge cases, boundary conditions, and failure cases.

Example function:

def calculate_discount(price, percentage):
return price * (percentage / 100)

Possible AI-generated scenarios:

• Normal discount case

• Zero discount

• 100% discount

• Negative percentage

• Very large price values

• Rounding behavior

Your job is to evaluate which tests reflect real system behavior.

AI suggests.
You decide.

3. Write the Tests Yourself

Do not copy-paste AI-generated test code.

Instead:

1. Review the AI test ideas

2. Select the meaningful ones

3. Write the tests manually

This preserves:

• understanding

• design clarity

• debugging ability

Example:

def test_zero_discount():
assert calculate_discount(100, 0) == 0

4. Compare Your Tests With AI Suggestions

After writing your tests:

Ask AI:

Compare these unit tests with your earlier suggestions.
What cases might still be missing?

This is where AI shines as a coverage reviewer.

You may discover:

• missing edge cases

• input validation gaps

• boundary conditions

5. Implement the Code to Pass Tests

Now return to the TDD loop.

Let the tests drive implementation.

AI can help with:

• implementation suggestions

• refactoring

• simplifying logic

• identifying duplicated code

Prompt example:

Given these tests, suggest a simple implementation that passes them.
Do not add features not required by the tests.

6. Use AI for Safe Refactoring

Once tests pass, AI can help identify design improvements.

Ask:

Refactor this code while preserving behavior verified by the tests.
Focus on readability and simplicity.

Your safety net:

The test suite.

If tests pass, refactoring is safe.

Exercise

Goal

Practice using AI to expand test coverage while maintaining developer control.

Step 1 — Pick a Small Function

Choose something simple:

• string parser

• calculation function

• validation logic

• utility method

Step 2 — Ask AI for Test Cases

Example prompt:

Generate unit test cases for this function.
Include edge cases and failure scenarios.

Step 3 — Write Tests Yourself

Do not copy the AI output.

Instead:

• read the suggestions

• select meaningful ones

• write tests manually

Step 4 — Compare Gaps

Ask AI:

Compare my tests with the earlier suggestions.
What important cases might still be missing?

Step 5 — Expand Coverage

Add the missing cases you agree with.

Your final test suite should reflect:

• real requirements

• edge conditions

• error behavior

Key Principle

AI improves test discovery.

Developers maintain design ownership.

A useful mental model:

You stay in the driver’s seat.