Thomson Reuters, the London-headquartered global media and information group, is running autonomous AI agents across its enterprise stack using Snowflake's CoCo platform — alongside US firms Fanatics and WHOOP — marking a global shift from AI assistants to systems that act independently at machine speed.1
CoCo functions as Snowflake's agentic control plane, giving enterprises a governed environment to manage workflows across data, models, and applications.1 The platform addresses a structural problem common to every major economy: existing technology was built for humans, not machines.
"Your existing tech stack was designed for human-operated, application-centric workflows. It needs to be reconsidered when the actor is an AI agent operating at machine speed across multiple systems simultaneously," said Surojit Chatterjee in MIT Technology Review.2
The architectural shift is not incremental. AI agents move across software layers, connecting systems and contextualizing data from multiple applications at once — a capability that breaks conventional IT governance models worldwide.2 Prasun Shah, writing in MIT Technology Review, identified this as the next competitive frontier for enterprises globally.2
On the services side, EXL — which operates delivery centers across India, the Philippines, and Eastern Europe — generated nearly $300 million in free cash flow in 2025 and is rebuilding its core business model around agentic delivery.3 Its investor day signalled a capital reallocation strategy aimed at AI-native operations at scale.
Hardware suppliers are aligning to the same timeline. Dell's exascale storage systems, NVIDIA GPU-accelerated data platforms, and AMD MI350P-equipped PowerEdge servers are all targeting H2 2026 availability — coordinated to support production-grade agentic deployments across global infrastructure.
Governance frameworks face equal pressure. When agents execute tasks autonomously across finance, compliance, and customer operations simultaneously, traditional accountability structures — designed for human decision-makers — collapse. Enterprises in every jurisdiction must now define who is responsible for agent outputs, not just human choices.
The transition from pilot to production is complete in leading organisations. Platforms are live, hardware timelines are confirmed, and enterprise capital is moving toward agentic infrastructure. The operating model redesign is global and underway.
