AI Coding Agents Produce Increased Code Volume Without Enhancing Software Development

The introduction of AI coding agents has brought about a significant shift in the software development landscape. Research indicates that the code review process has seen an average delay of 49 percent following the deployment of these AI tools. This increase in review time is reflected in detailed metrics, with the incidence of pull requests requiring changes nearly doubling and comments per pull request rising by 35 percent after AI agents were integrated into workflows.

In light of this shift, the study reveals a 14 percent uptick in the proportion of employees involved in code reviews after the introduction of AI agents. However, the researchers note that they couldn’t link these changes to significant job losses within organizations, as they examined overall employment levels at companies like Jellyfish and cross-verified this data with LinkedIn statistics.

Pull requests need changes a lot more often in the “agentic coding” era.

While the potential benefits of AI in facilitating code reviews are theoretically promising, the current research indicates that its influence remains limited. By March 2026, around 80 percent of the companies surveyed had adopted some form of AI code review system. However, AI agents accounted for only 23.3 percent of all review comments and 10.8 percent of all pull requests, underscoring the continued reliance on human expertise.

Given that AI coding agents are still evolving, firms are likely navigating a learning curve in optimizing their use. Around 95 percent of the organizations involved in the study have integrated these agents into their operations, yet many are still figuring out when and how to employ them effectively. As engineering teams gain experience, the balance between coding speed and review time may improve, reflecting a better understanding of the strengths and limitations of AI.

In the current landscape, the advantages of AI-assisted coding appear to be counterbalanced by the increased demands placed on human reviewers. This paradox raises questions about whether the investment in AI tools is justified for many firms, considering both time and financial implications.