This study, conducted in collaboration with scholars from Harvard Business School, MIT Sloan School of Management, the Wharton School at the University of Pennsylvania, and the University of Warwick, investigates the impact of generative AI (Gen AI) on the productivity of highly skilled workers. Focusing on over 750 BCG consultants globally who served as participants, the study examined the degree to which Gen AI (specifically OpenAI’s GPT-4) influenced the participants’ performance on professional services tasks mirroring their daily work activities. Key findings include: 1) When using OpenAI’s GPT-4 for creative product innovation— a task involving ideation and content creation— around 90% of participants improved their performance. Furthermore, the improvement was 40% higher than those working on the same task without GPT-4. Workers best captured this upside when they did not attempt to improve the output that the technology generated. 2) However, for business problem-solving tasks, using GPT-4 resulted in performance that was 23% lower than that of the control group (see Exhibit 1). These outcomes underscore the necessity for thoughtful consideration of the compatibility of Gen AI with different organizational tasks. The authors also propose that organizations establish a “generative AI lab” to experiment with various use cases for their organizations. In case you missed it last week, here is a newly released 26-page report from MIT Technology Review Insights on how organizations are experimenting with and deploying Gen AI.