AI coding tools increase task output but drive 91% rise in review time

The integration of generative AI into software development is creating a significant shift in engineering workflows. While AI-powered coding tools have increased the number of completed tasks by approximately 21 percent, they have also led to a 91 percent increase in the time required for code reviews.

Industry experts and reports from Bain & Co. suggest that the primary bottleneck has shifted from writing code to verifying its accuracy. Developers are increasingly moving from active design roles to a supervisory role, often struggling to fully comprehend or debug AI-generated code. This creates risks regarding technical debt, hallucinations, and a potential decline in core programming competencies.

In the agency sector, AI adoption is rapidly expanding into support functions like finance and HR, with usage in these areas rising from 7 percent to 34 percent over a six-month period. While AI is being used to increase service volume and efficiency, its role in creating entirely new products has seen a slight decline.