"Maybe We Need Some More Examples:" Individual and Team Drivers of Developer GenAI Tool Use
Despite the widespread availability of generative AI tools in software engineering, developer adoption remains uneven. This unevenness is problematic because it hampers productivity efforts, frustrates management’s expectations, and creates uncertainty around the future roles of developers.
With the goal of understanding what differentiates developers who are frequent versus infrequent GenAI tool users within the same team contexts we utilize a paired interview design, conducting sequential semi-structured interviews with 54 developers representing 27 pairs from the same team matched on key factors- primary programming language, role, and seniority- but who exhibit contrasting usage patterns (one frequent and one infrequent user).
We identified key differences in approaches frequent and infrequent users including how they perceive the tool (as a collaborator vs. feature), their engagement approach (experimental vs. conservative), and how they respond when encountering challenges (with adaptive persistence vs. quick abandonment).
Our analysis also identified key commonalities between the two groups: for many developers, organizational factors can actively shape these individual factors through an amplification effect, e.g., team-specific demonstrations of applying GenAI tools on common development tasks can transform developers’ perception of tool usefulness
Most critically, we identified a phenomenon which we refer to as the Productivity Pressure Paradox: increased productivity expectations from management without corresponding support often create a paradoxical effect, where developers lack the time necessary to develop the skills that would save time. This finding challenges the prevailing GenAI tool deployment strategy across the software industry, which frames the challenge of determining how to use these tools to yield the expected productivity gains as the responsibility of individual developers. Like historical technological transformations in manufacturing and software engineering, we argue systematic productivity gains require systematic organizational change, not individual heroics.
