GitHub Copilot leads as AI coding tools surge in adoption among engineers

Artificial intelligence-powered coding tools have become nearly ubiquitous among software engineers, according to a recent survey by Jellyfish, a company that specializes in developer management software. In May, Jellyfish surveyed 645 full-time engineers, managers, and executives from organizations of varying sizes, revealing that 90% of engineering teams now incorporate artificial intelligence solutions into their workflows—a figure that has jumped from 61% just a year ago. Nearly a third of respondents reported their organizations have formally adopted artificial intelligence tools, while another 39% are currently experimenting with them. Only 3% remain uninterested in artificial intelligence for coding and do not plan to change course.

The data also highlights a diversified approach to adoption: 48% of respondents said they rely on two or more artificial intelligence coding tools instead of consolidating on a single platform. The survey focused specifically on coding tools designed for engineering, excluding general-purpose chatbots such as ChatGPT, in order to track platforms purpose-built for software development use cases. Among these, Microsoft’s GitHub Copilot leads the field, with 42% of engineers naming it their tool of choice. Google’s Gemini Code Assist secured a strong second place, while Amazon Q (the rebranded CodeWhisperer) and Cursor tied for third, together forming a dominant group in the burgeoning artificial intelligence code assistant market.

Beyond tool preference, the Jellyfish report measures real-world impact: 62% of engineers said that artificial intelligence tools have boosted their productivity by at least 25%, and 8% reported a doubling of their output. Fewer than 1% felt artificial intelligence was slowing their work. Looking forward, 81% of respondents expect at least a quarter of current engineering workloads will be automated by artificial intelligence within five years—but most envision ongoing collaboration between engineers and artificial intelligence systems rather than full automation, reflecting a human-machine hybrid approach to creativity and problem-solving in software development.