AI agents produce code faster than any team can review. That's a fact. It's already happening, in BC world too. And code review, the safety net we've relied on for years, was never as effective as we kept telling ourselves it was.

But here's the thing: test-driven development (TDD), the workflow that in theory we always knew was right, but never had the discipline to sustain, turns out to be a natural fit for agents. They don't mind writing tests first. They don't get bored of red-green-refactor. They don't cut corners when a deadline looms.

But there is a catch, of course. Just like us, agents know TDD is good in theory, but they won't apply it unless you force them. In this session I'll show you how I set this up in Business Central: how to structure code for testability, how to teach the agent a strict test-first cycle, and how to use mutation testing to prove your tests actually mean something.

No human can meaningfully review thousands of lines of generated business logic. But anyone can read a test and know exactly what it proves.