and only add human intervention where automation can’t deliver or where you don’t want it to, for reasons like creative nuance or brand tone. In this big shift, humans move from being the doers to the stewards and curators of automated systems, which is a major mindset change. So, going back to first principles, challenging legacy processes, and being prepared to lead cultural transformation is what drives adoption. The tech will catch up, but your ability to adapt how you work is what will unlock real value. Jochen Hartmann: Our advice is to focus on effectiveness rather than efficiency. Efficiency gains, especially in content creation, are relatively straightforward. Driving effectiveness uplifts is key. We also found that templates foster creativity rather than limit innovation—they enable consistent quality at scale and help build trust in the AI’s outputs. And it’s important to assess how gen AI can disrupt end-to-end workflows to materialize disruptive efficiency gains, not just incremental ones. That’s also important when designing agentic AI systems and workflows. Organizations should start with systematic experimentation—rigorous field testing beats theoretical planning. Collect data and use it to inform predictive AI systems (such as the personalization engine of MAGE); measure what matters by focusing on engagement and effectiveness metrics, not just efficiency gains; and build organizational buy-in, involving stakeholders in the development process to overcome natural skepticism and have top management lead the way. McKinsey: Which domains/use cases are next on your gen AI road map and why? Thomas Coesfeld: We’ve identified the next wave of gen AI use cases based on a combination of strategic value, technical feasibility, and speed to impact. Following our initial rollout of seven use cases across investment, finance, marketing, technology and operations, B2B, and legal, we’re now focused on scaling into high-impact domains, particularly where automation, insight generation, or personalization at scale can unlock major gains in effectiveness or revenue gains. These next-wave priorities reflect a shift from isolated automation to full-stack, cross-functional enablement. Ultimately, our road map aims to embed gen AI across the business, enabling every team to operate faster, smarter, and at a radically larger scale. Thomas Coesfeld is CEO of BMG, the music division of Bertelsmann. He was previously BMG’s CFO and held leadership positions within Bertelsmann. Jochen Hartmann is a professor of digital marketing at the TUM School of Management, Technical University of Munich, focusing on machine learning, social media, multimodal digital advertising, and human–machine interaction. Comments and opinions expressed by interviewees are their own and do not represent or reflect the opinions, policies, or positions of McKinsey & Company or have its endorsement. 47 Past forward: The modern rethinking of marketing’s core