AI Speeds Up Coding but Leaves Software Complexity Behind

AI coding tools can help teams write software faster, but they do not remove the work needed to make it reliable. In enterprise development, code still needs quality assurance, security reviews, regulatory checks and integration with existing systems.

In a 2025 study, Tilburg University researchers analyzed thousands of open-source projects. They found that AI coding tools increased output but also added to senior developers’ review workload. Experienced developers spent significantly more time reviewing AI-generated code, while their own coding output fell by nearly 20%.

The article argues that smaller teams of two or three people can now explore ideas that might once have required larger teams, more specialized skills and more time. But producing code is only part of the job: organizations also need to validate, operate, maintain and evolve the resulting systems.

Each new component can bring additional needs for testing, security, integration, monitoring and long-term maintenance. The author, NiCE’s General Manager of CX Engineering, says the challenge is managing the complexity that faster code generation creates. A prototype still needs validation, security and reliability work before it can be treated as a finished product.

The article was published as part of TechRadar Pro Perspectives. The views are the author’s and do not necessarily represent TechRadar Pro or Future plc.