AI-Native Software Engineering

Chapter 8. Verification and Quality Gates

A Note for Early Release Readers

With Early Release ebooks, you get books in their earliest form—the author’s raw and unedited content as they write—so you can take advantage of these technologies long before the official release of these titles.

This will be the 8th chapter of the final book. Please note that the GitHub repo will be made active later on.

If you’d like to be actively involved in reviewing and commenting on this draft, please reach out to the editor at [e-mail].

The last few chapters were all about steering an agent toward the right answer before it writes a single line of code. Context engineering gave you ways to load the agent with what it needs: your conventions, your architecture decisions, the rules you actually care about. SDD showed you how to pin down intent upfront, so the agent builds the thing you meant it to be build, instead of a plausible-sounding wrong thing. That’s all about inputs. It answers one question: How do you guide the agent?

This chapter takes up the question that comes next. The agent has already run. It’s read your spec, it’s written the code, and it is sitting on a branch ready to open a pull request. The code looks fine. But how does it know the code is correct? And how do you know, with 20 more pull requests waiting behind this one?

The Bottleneck Moves

For most of the history of software teams, the slowest step was writing the code. An engineer drafted it, ran it in their head, ...