Financial institutions are moving from isolated agentic AI experiments to multi-agent AI systems across banking and insurance workflows, but only 10% have achieved enterprise-wide orchestration. Capgemini’s World Cloud Report for Financial Services 2027 says 59% are piloting or deploying such systems, while 90% use or plan to implement intelligent model routing (directing tasks to suitable models) to manage performance and token costs. Eight in 10 executives prioritize reducing tech debt, and 52% of banks and insurers prefer hyperscalers (large cloud providers) to govern deployments. By 2028, executives prioritizing business outcomes as the main AI-funding criterion are projected to rise 141%, while efficiency-focused executives are projected to fall 42%. Three-quarters expect AI agents to become important to customer engagement within three years. Hybrid sourcing is expected to become the dominant approach for more than two-thirds of institutions within three years, while only 5% have already institutionalized agentic AI at enterprise scale.