The Challenge
While human-computer interaction and software engineering scholarship have extensively explored developer interactions with Generative AI (GenAI), critical gaps persist regarding cross-functional roles like Product Managers (PMs). Product managers occupy a distinct socio-technical nexus, balancing technical execution, human empathy, and strategic business goals. However, the theoretical mechanisms governing task delegation—the proactive transfer of authority versus mere operational usage—remain poorly defined. Prior research overlooks how knowledge workers systematically evaluate task risks, establish agency boundaries, and navigate the 'jagged technological frontier' without surrendering human accountability to autonomous algorithmic systems.
Core Findings
Conducting an exploratory sequential mixed-methods study at Microsoft, the authors surveyed 885 PMs, analyzed telemetry data across 731 workers, and interviewed 15 practitioners. The study establishes a 'Selective Delegation Framework' articulating how PMs deliberate task handover across individual, team, and organizational dimensions. PMs readily offload operational writing, summarizing, and ideation, yet deliberately withhold high-stakes strategic synthesis, ethical oversight, and stakeholder management. Critically, the authors delineate delegation (transferring decision rights and accountability) from mere tool usage. Findings reveal that while GenAI accelerates workflows, PMs strictly enforce an ethical paradigm: accountability must never be transferred to non-human actors. Concurrently, the boundaries between software engineering and product management roles are progressively blurring.
Strategic Takeaway
For technology executives and digital leaders, this study provides an actionable framework to guide corporate AI governance. AI adoption models cannot simply measure usage frequency; organizations must establish clear boundaries defining delegable versus non-delegable tasks. Leaders should implement guidelines reinforcing non-transferable human accountability, preventing algorithmic complacency in strategic roadmapping. Furthermore, executive development must shift toward training knowledge workers in critical evaluation, prompt orchestration, and meta-cognitive oversight, ensuring that technical boundary blur between engineers and product leaders enhances cross-functional synergy rather than generating organizational ambiguity.