Automated review

Also called: AI code review, AI review, review agent, automated code review, bot review

Having an AI model read and critique work, usually code, before a person does. It catches problems that rule-based checks miss, such as confusing logic or a forgotten requirement, but it can be wrong and does not replace a person signing off.

For example: Each change the agent proposes is first read by a separate reviewing agent, which flags a missing error message before the founder looks.