The SPACE of AI: Real-world lessons on AI’s impact
Developers overwhelmingly believe AI enhances their productivity, yet its impact is far from uniform across all tasks and teams.
Welcome to the latest issue of Engineering Enablement, a weekly newsletter sharing research and perspectives on developer productivity.
Take our survey about the effect of AI on perceived productivity:
Do you think managers and developers agree on the productivity gains from AI? If you’re a developer or a manager of developers, you can help us answer this question. Go here to participate in our 15-minute survey to share how you think AI is changing developer productivity.
This week I read The SPACE of AI: Real-World Lessons on AI's Impact on Developers, a recently published paper by Microsoft researchers that studied how AI is impacting developer productivity and satisfaction across the five SPACE dimensions.
My summary of the paper
Most studies on AI’s impact take a narrow view, measuring only how much faster developers can complete coding tasks. But productivity is multidimensional, and coding is just one aspect. In fact, developers spend only 14% of their time writing new features, making it clear how incomplete a picture it is to focus on coding alone.
This study combined survey data (including 530 responses, primarily from Microsoft but also including other companies like Airbnb, Meta, and Netflix) with interviews and observational studies. The surveys aimed to assess how AI tools impact developers’ experiences across the SPACE dimensions. Additionally, respondents were asked about team-level adoption, the availability of AI training resources, and their own frequency of AI use.
Here’s what the study found:
Adoption: How widely are developers using these tools?
The study started by asking developers how often they use AI—this way, researchers could later identify whether the frequency of usage plays a role in AI’s impact on productivity. This part of the study yielded some interesting insights:
Most developers regularly use AI. 75% of developers said they regularly use AI to complete tasks. Additionally, among those who have adopted AI, 64% report using it at least once a week.
Seniority has little impact on adoption: Developers with 7+ years of experience are only 4% less likely to be daily AI users than those in their first three years.
Organizational support drives adoption: Developers at companies that actively encourage AI use are seven times more likely to be daily users than those at companies that do not.
AI and productivity: Where are the gains?
The researchers then turned to the central questions: Are AI tools making developers more productive? And do some aspects of work benefit more than others? To find out, they examined developers’ perceived productivity gains across the five SPACE dimensions.
Developers who regularly use AI report strong benefits: 90% say it makes them more productive, and 80% say they would be disappointed if they could no longer use it. Fewer than 3% disagreed that AI had a positive impact in any SPACE dimension.
We can see from the table above that developers report the strongest productivity gains from AI in task throughput and efficiency. Additionally, 71% of AI adopters believe these tools help them deliver customer value, and 62% report higher job satisfaction with AI. These results suggest that, for those who use AI, it’s not just making them faster but also helping them feel more effective and fulfilled.
Collaboration, on the other hand, tells a more nuanced story. Fewer than half (48%) of adopters feel AI improves their ability to collaborate with teammates. Most neither agreed nor disagreed, showing that AI’s role in this area is still evolving.
The follow-up interviews added helpful context. AI doesn’t negatively impact collaboration but rather reshapes it. Managers observed fewer interruptions, as developers relied less on colleagues for quick coding answers. One executive respondent noted that AI reduced the “reputational damage” developers sometimes felt when asking teammates basic questions.
Factors influencing AI’s impact on productivity
The study found that AI’s impact depends on several factors: the type of task, a developer’s familiarity with the tools, how often it is used, and the extent of team-wide adoption.
Task type and complexity. Developers report that AI excels at routine, repetitive work but struggles with complex or novel challenges.
Developer skill and familiarity. The more developers use AI, the better they learn when and how to apply it. Early interactions with AI might be limited to basic code suggestions, but over time, many expand to problem-solving, documentation, and debugging. As familiarity grows, so does perceived value. Frequent users tend to report higher productivity gains.
Frequency of use. A positive link exists between how often developers use AI and how strongly they believe it improves productivity. But causation is unclear. Smaller sample sizes among infrequent users make the differences statistically uncertain, and it remains unclear whether frequent use drives productivity or whether those who see value are simply more inclined to use it.
Team-wide adoption. Adoption at the team level amplifies impact. Shared best practices, cultural norms, and seeing peers benefit from AI all shape how individuals perceive AI’s value. The more teammates who use AI, the more productive the team appears. This dynamic also raises the possibility of social pressure: as AI becomes the norm, those who don’t use it may be seen, or may feel, as less productive.
Final thoughts
This paper offers a more comprehensive view of AI’s impact, going beyond the question of whether it simply helps developers complete tasks faster. What stood out most to me were the factors influencing that impact. The study supports the idea that organizational support and team norms play a big role: developers at companies that actively encourage AI use are seven times more likely to be daily users than those without support. The findings also reinforced that task type matters. AI is highly effective for routine, repetitive work, but it struggles with complex or novel challenges.
Who’s hiring right now
This week’s featured DevProd & Platform job openings. See more open roles here.
Deliveroo is hiring a Staff Platform Product Manager | London, UK
Rippling is hiring a Director, Platform Engineering | San Francisco, CA
Atlassian Williams Racing is hiring a Software Engineer - Engineering Acceleration | Grove, UK
RH is hiring a Director, Platform Engineering | Pleasanton, CA
ScalePad is hiring a Head of AI Engineering & Enablement | Canada (Remote or in-office)
That’s it for this week. Thanks for reading.
-Abi