How Companies Organize the Deployment of Generative AI – and Who Is Responsible

Our survey analysis shows that CEO oversight of AI governance—meaning the policies, processes, and technologies required for the responsible development and deployment of AI systems—is one of the elements most strongly associated with the reported bottom-line impacts of an organization’s use of generative AI. This is especially true in larger companies, where the CEO’s influence on AI is greater. Twenty-eight percent of respondents whose organizations use AI reported that their CEO is responsible for overseeing AI governance, although this share is lower in larger organizations with annual revenues of $500 million or more, where 17 percent say AI governance is overseen by their board of directors. In many cases, AI governance is jointly held: on average, respondents report that two leaders are responsible.

The value of AI comes from rewiring how companies operate, and the latest survey shows that among 25 tested organizational features across all sizes, workflow redesign has the greatest impact on an organization’s ability to see EBIT effects from generative AI use. Organizations are beginning to reshape their workflows as they deploy GenAI. Twenty-one percent of respondents whose organizations report using generative AI say their organizations have fundamentally redesigned at least some workflows.

Tariffs and Global Trade: Examining the Impact on Commerce Join McKinsey senior partners Cindy Levy and Shubham Singhal to discuss how business leaders can gain advantage by identifying short- and long-term strategic actions in response to rapid changes in trade policy.

Organizations have employees who monitor the quality of generative AI outputs, though the extent of this oversight varies widely. Twenty-seven percent of respondents whose organizations use generative AI say that employees review all AI-generated content before it is used—for example, before a customer sees a chatbot response or before an AI-generated image is used in marketing materials (Figure 2). A similar share say that 20 percent or less of GenAI-generated content is reviewed before use. Respondents working in business services, legal, and other professional services are much more likely than those in other industries to say that all outputs are reviewed.