In most organizations, there is a limited number of good code reviewers. This is inconsistent, unscalable, and easy to skip under deadline pressure or when reviewers are unavailable. In this talk, I share the concrete 39-rule review checklist my team built by turning every recurring test smell into an explicit, citable rule, and how we paired it with an AI agent reviewer that runs the full rule matrix on every pull request, consistently.
This talk has been presented at AI Coding Summit Asia 2027, check out the latest edition of this Tech Conference.





















