General Motors redesigned its engineering workflows around AI agents in its autonomous vehicle division, tripling the rate of merged pull requests and accelerating development cycles.
Software engineers at GM's autonomous driving unit now spend only 15 percent of their time writing code. AI agents handle the remaining 85 percent of engineering work, including analyzing vehicle data, triaging problems, running experiments, and testing potential fixes. Rashed Haq, GM's VP of autonomous vehicles, described these results at VB Transform 2026.
The impact is substantial. GM's autonomous vehicle engineering organization now produces roughly three times as many merged pull requests compared to previous workflows. The shift also shortened release cycles and reduced defects that escape into later development stages.
GM's approach reflects a broader industry shift toward using AI agents not as replacements for engineers but as force multipliers for repetitive, time-consuming tasks. By automating the work between code writing and deployment, GM freed engineers to focus on higher-level problem-solving and architectural decisions.
The engineering workflow redesign required more than just plugging in AI tools. GM restructured how teams collaborate, how code reviews happen, and how problems are prioritized. AI agents now serve as the first line of triage, catching issues before human engineers see them.
This model has direct implications for autonomous vehicle development timelines. Faster iteration means more testing cycles and faster validation of safety-critical systems. For a company racing to deploy self-driving technology, acceleration in the development pipeline translates to competitive advantage.
The numbers Haq shared suggest the approach works at scale. Three times as many merged pull requests does not mean three times as many features shipped, but it does indicate substantially higher velocity in experimentation and bug fixing. In software development, throughput matters when you're trying to solve complex problems like autonomous driving safety and reliability.
GM's success could prompt other automotive and tech companies to restructure their engineering