While artificial intelligence (AI) is seen as the ultimate productivity tool for software development, the impact of AI development tools on teams is still in its infancy.
These are takeaways from a recent report on DevOps trends, which is published by Google Cloud’s DevOps Research and Assessment (DORA) team, and based on data from 36,000 technology professionals worldwide.
Also: Implementing AI into software engineering? Here’s everything you need to know
It’s common for experts to suggest these days that AI will deliver significant boosts to software development and deployment productivity, along with developer job satisfaction.
“So far our survey evidence doesn’t support this,” the report’s authors, Derek DeBellis and Nathen Harvey, both with Google, state.
Also: Is AI in software engineering reaching an ‘Oppenheimer moment’?
“Our evidence suggests that AI slightly improves individual well-being measures — such as burnout and job satisfaction — but has a neutral or perhaps negative effect on group-level outcomes such as team performance and software delivery performance.”
These flat findings are likely due to the fact that we’re still at the early stages of AI adoption, they surmise: “There is a lot of enthusiasm about the potential of AI development tools, as demonstrated by the majority of people incorporating at least some AI into the tasks we asked about. But we anticipate that it will take some time for AI-powered tools to come into widespread and coordinated use in the industry.”
Also: Six skills you need to become an AI prompt engineer
Despite the limited impact of AI so far, the survey does identify the factors that are moving development shops forward. In their research, DeBellis and Harvey isolated a segment of “elite” professionals who are on top of their game. These professionals only require lead times of one day to make changes in applications, versus a week to a month at low-preforming shops. They can deploy software multiple times a day. They also report change-failure rates for buggy software of 5% or less. By contrast, those in low-preforming software shops have rates exceeding 60%.
While AI may help IT professionals in the future, there are best practices that the elite group is pursuing that are making a difference today. The co-authors identify those practices:
- Build with users in mind: The Google research shows “that a user-centric approach to building applications and services is one of the strongest predictors of overall organizational performance. Teams that focus on the user have 40% higher organizational performance than teams that don’t.”
- Establish a healthy culture: “Teams with generative cultures, composed of people who felt included and like they belonged on their team, have 30% higher organizational performance than organizations without a generative culture.”
- Strive for high-quality documentation: “High-quality documentation amplifies the impact that DevOps technical capabilities — for example, continuous integration and trunk-based development — have on organizational performance. Overall, high-quality documentation leads to 25% higher team performance relative to low-quality documentation.”
- Distribute work fairly: “We find that respondents who take on more repetitive work are more likely to experience higher levels of burnout, and women and members of underrepresented groups are more likely to take on more repetitive work. Women or those who self-reported their gender do 40% more repetitive work than men.”
- Leverage cloud flexibility: “Using a public cloud, for example, leads to a 22% increase in infrastructure flexibility relative to not using the cloud. This flexibility, in turn, leads to teams with 30% higher organizational performance than those with inflexible infrastructures.”
Contrary to widespread and deeply ingrained impressions, software developers do not work in isolation. Instead, they work in teams, and strive to focus on their business. The survey helps shed light on what’s important for top-performing developers — and AI is still more of a shiny object than a differentiator.
