Using Generative AI for Automated Testing

What if you could offload most of your testing work to AI—without sacrificing quality?

In this talk, I’ll walk through how I use generative AI and LLMs to fundamentally change how I build and test software. I’ve moved away from traditional Test-Driven Development (TDD) toward a more behavior-driven approach where testing is continuously embedded throughout development.

Instead of writing tests first, I collaborate with AI agents that help identify edge cases, validate expected behavior, review code, and surface potential security and accessibility issues in real time.

To make this concrete, we’ll build part of a restaurant point-of-sale system together. Servers can manage tables, create orders, and track timing—but we’ll focus on implementing the payment flow:

• Selecting a table
• Viewing the current tab
• Processing a payment

Along the way, I’ll demonstrate:

• How I structure prompts for different AI “agents”
• Using AI for automated QA and regression testing
• Catching edge cases and unexpected behaviors early
• Performing lightweight security and accessibility reviews
• Treating AI as a continuous peer reviewer

This session is practical, opinionated, and geared toward developers who want to rethink testing in the age of AI.