Six Hard Questions Enterprises Must Now Answer in the Age of Agentic AI

At KPMG's annual Tech and Innovation Symposium, six foundational questions emerged that enterprises are grappling with as AI shifts from an assistive tool to one that autonomously performs work. KPMG's Steve Chase warned that bolting AI onto existing processes — workable in the assisted-AI era — becomes far more costly and risky as agentic AI takes hold. Companies are now being pushed to think in terms of multi-model architectures, cost visibility at the token level, and deliberate workforce reskilling rather than relying on training videos alone. The consensus was that AI adoption requires genuine organizational redesign, with early adopters pairing directly with business units to spread practical knowledge. Beyond internal transformation, firms are also beginning to rethink external business models, including shifting from input-based to outcomes-based pricing and reconsidering what their core products even mean in an agent-driven world.