Large-scale deployments of “agentic AI” in enterprise environments are beginning to take shape, with PwC unveiling more than 120 production-grade AI agents built with Google Cloud. The portfolio spans 24 cross-functional workflows and the global consultancy says it shows agentic AI is moving beyond the pilot stage toward fully governed, integrated systems.
The agents leverage Google Cloud technologies including Vertex AI, Gemini models, and the Agent2Agent (A2A) protocol, alongside PwC’s orchestration tools. According to the company, the systems are architected to integrate large language model reasoning, domain-specific knowledge and real-world data, with early client deployments showing up to eight times faster cycle times and reducing costs more than 30 percent in some cases.
“Moving from proof-of-concept to production at this scale marks a turning point,” said Jason Ruge, Google Cloud alliance leader and practice principal for PwC US. “These 120 agents are designed for production use—each one is integrated, governed and auditable. This is intelligent scale: the ability to spin up value-driven agents and workflows on demand, securely and at enterprise-grade. And, as part of PwC’s agent OS, they enable AI transformation across industries and business functions.”
The company said it has launched an early access program offering multimodal agents, industry-specific bundles, and a plug-and-play agent marketplace for accelerated deployment. While latency and orchestration overhead remain challenges, PwC says ongoing development with Google Cloud’s Agent Development Kit and Agent Engine is aimed at improving performance and scalability.