“Vibe‑coding” is trending: prompt until something seems to work. The outcome is often code that's difficult to understand and maintain.

In this session, the speakers introduce EXACT Coding — Example‑Guided, AI‑Collaborative & Test‑Driven — a pragmatic workflow for AI-assisted development with an emphasis on readability and maintainability. They begin by using Example Mapping to clarify terminology and boundaries, translating examples into clear, meaningful tests. Rule files and explicit commands structure iterative prompting, letting teams guide the AI intentionally. Tests serve as a living specification, supporting systematic refactoring and helping keep the code base easily readable and maintainable. Throughout the process, human‑in‑the‑loop decisions remain essential: developers continually assess clarity, make structural improvements, and determine when to refine prompts.

A focused live coding demo on a representative domain showcases this workflow end to end. Attendees will gain concrete strategies for making code more readable, reducing accidental complexity, and ensuring maintainable solutions — regardless of technology stack — so they can leverage AI productively without drifting into an unreadable codebase.