Irina von Rosén

Agentic AI is exposing the limits of traditional enterprise structures. Siloed teams, slow decision-making, and fragmented data foundations cannot support autonomous, real-time systems. As organisations move from experimentation to scaled deployment, the operating model, not the technology, is emerging as the primary constraint. This panel convenes leaders who are fundamentally redesigning how their organisations operate to enable agentic AI at scale, from global AI hubs to new cross-functional ways of working.

• Redesign operating models to support real-time, agent-driven decision-making across the enterprise
• Balance centralised AI expertise (e.g. hubs/CoEs) with embedded, domain-level ownership
• Break down silos between data, engineering, and business teams to enable end-to-end execution
• Drive organisation-wide adoption and agentic literacy, not just isolated technical capability
• Align governance, talent, and workflows to support continuous scaling, not one-off deployments 

As organisations scale AI, they face an increasingly complex regulatory landscape shaped by the EU AI Act, the Data Act, the Data Governance Act, and GDPR. Together, these frameworks are redefining how AI systems are built, how data can be used, and who ultimately holds control. The stakes are rising quickly. Under the EU AI Act, non-compliance with prohibited practices could result in fines of up to €35M or 7% of global annual turnover, making AI governance not just a legal concern, but a critical business risk.

As organisations accelerate investment in agentic AI, the conversation is shifting from potential to performance. While early use cases promise efficiency and transformation, many are discovering that the true cost of building, deploying, and scaling these systems is far more complex than anticipated. From token usage and infrastructure to integration, maintenance, and governance, costs accumulate quickly, often without a clear understanding of how value is measured or realised. At the same time, many organisations are still struggling to define what "ROI" actually means for agentic AI, or how it compares to simpler alternatives like automation or traditional machine learning. This panel brings together leaders to share a more transparent view of what it really takes to make agentic AI deliver measurable business impact.