Engineers using AI agents to generate more pull requests are creating QA bottlenecks, with teams responding through code review and manual testing requirements.

At one company, engineers are producing a high quantity of pull requests with the help of AI agents, while QA was already the bottleneck due to software involving camera flows across multiple devices.

One response has been to slow down engineers by adding multiple code reviews and encouraging engineers to engage in manual testing themselves to better understand the product.

A commenter notes that making pull requests faster does not help if QA remains slow, as the problem simply moves downstream to the QA team.