Gemini at Work: Knowledge Workers' Perceptions and Assessment of Productivity Gains

The rise of Generative AI (GenAI) presents a paradigm shift in knowledge work. To examine its impact on productivity, we conducted seven focus groups (n=37) with employees across diverse job functions in an enterprise setting, where workers engaged with a large language model (LLM) embedded in a suite of productivity applications. Our study identified four categories of GenAI-facilitated work activities: information management, content generation, problem solving, and communication and collaboration. These findings offer a grounded framework of GenAI-enabled practices for both researchers and practitioners, while also surfacing key challenges in realizing GenAI's promised productivity gains. Beyond conventional metrics like time savings or output volume, participants attributed productivity improvements to time redistribution, enhanced decision-making, and reduced reliance on traditional intermediaries. We contribute actionable insights for designing GenAI systems that support context-aware productivity in the evolving landscape of knowledge work.