AI and the Future of Refactoring

AI promised to solve the endless chore of refactoring, but its tendency to deliver massive, uncontrolled changes often creates more problems than it solves. This talk introduces Test-Driven Navigation (TDN), a practical workflow to tame your AI assistant. By providing failing tests as hyper-specific instructions, we can force AI to work in safe, incremental steps. Learn how to combine principles from Extreme Programming (the 3 C's) with the power of modern AI to keep your code continuously working, transform your refactoring process, and elevate your role from a hands-on coder to a strategic system architect.

Value for the audience:
The audience will learn how to use finely tuned, finely tailored AI assisted software development workflow to produce quality solutions that align with business values. The method, invented by the presenter, is called "Test Driven Navigation".

Problems addressed:
Majority of software development efforts tend to go toward maintaining the existing code. This presentation explains how to organize the workflow to minimize the maintenance efforts.

When using AI-assisted code creation tools, there is often a problem with unnecessary bloat, excessive accidental complexity which leads to quick accumulation of technical debt.

AI agents are notorious for creating more code than is necessary, violating the lean principles. The presenter demonstrates how ot use mutation testing to minimize the harmful bloat.

Talk language: English
Level: Advanced
Target group: Experienced technical leads/coach prctitioners

Company:
Digital Exprt Inc.

Alex Bunardzic