From Experiments to Enterprise AI Agents at Scale
Enterprise AI agents are software agents powered by artificial intelligence that can act on behalf of employees or systems across business workflows, using secure access to organizational data, tools and applications to automate tasks, support decisions and coordinate work at scale. After years of pilot projects, the market is now moving toward structured AI agent deployment that is secure, governed and aligned with enterprise operations. Two large consulting and technology services groups, KPMG and Atos, are showing what this transition looks like in practice. Both are rolling out Microsoft 365 Copilot across their workforces while adopting Microsoft Agent 365 as a control plane for their growing agent populations. Their strategies highlight what many large organizations now demand: consistent governance, clear ownership of AI agents, and tight integration with existing productivity and security platforms.
KPMG: Turning Copilot and Agent 365 into an Enterprise Playbook
KPMG is extending its use of enterprise AI agents on two fronts: internally for its 276,000 professionals and externally for audit, tax and advisory clients. Two years after its first Microsoft 365 Copilot deployment, KPMG member firms are expanding access across the global workforce, embedding AI into everyday tasks such as document drafting, analysis and collaboration. In parallel, KPMG will adopt Microsoft Agent 365 to manage how AI agents are deployed, monitored and updated across its organization, integrating it into the KPMG Workbench ecosystem and Trusted AI framework. This brings centralized governance, so AI agents working across different systems and data sources can be observed and controlled from one place. According to KPMG’s Global Chief Digital Officer Lisa Heneghan, this expansion is about “embedding responsible AI into the heart of our culture and helping clients do the same.”
How KPMG Helps Clients Move from Pilots to Production
KPMG’s approach to enterprise AI agents is designed to move clients beyond isolated proof-of-concept projects toward systematic AI agent deployment. Member firms are working with Microsoft to help organizations put AI agents into production with governance, security and controls in place, applying the same principles KPMG uses internally. This includes integrating agents across systems and workflows, not just within a single application, so they can drive measurable business outcomes such as faster reporting, streamlined regulatory work or more responsive supply chains. Clients receive guidance on governance, risk and compliance frameworks, including lifecycle management and clear ownership for each agent. For example, Integra LifeSciences is embedding Copilot into core functions like Global Supply Chain and Regulatory Affairs as part of an enterprise AI operating model, tracking adoption and return on investment to ensure reliable, repeatable results.
Atos: Workforce-Wide Agentic AI Workforce Deployment
Atos is taking a workforce-wide approach to agentic AI, rolling out Microsoft 365 Copilot to all 56,000 employees across 54 countries and using Microsoft 365 E7 as the foundation. E7 brings together Microsoft 365 E5 security and compliance, Copilot and Agent 365, augmented with Entra Suite, Defender, Intune and Purview. Atos is already managing a fast-growing population of 19,000 AI agents through Agent 365, covering agents that act on behalf of users and those that operate with their own credentials. This means Atos IT and security teams can govern AI agents using familiar admin and security workflows. The company is also deploying Microsoft Copilot Studio and Microsoft Foundry to design and operate agents for internal IT, business functions and client projects, under a single operating model that feeds into its Sovereign Agentic AI studios for bringing AI safely into production.
Secure, Trusted Agentic AI as a Competitive Advantage
Across both KPMG and Atos, a pattern emerges for enterprise AI agents: success depends less on isolated models and more on secure, governed platforms that can scale. Their deployments show that Copilot enterprise adoption is tightly coupled with Agent 365, which provides the visibility and control needed when thousands of agents act inside complex organizations. Security and compliance tools such as Entra, Defender, Intune and Purview are not add-ons but core to the operating model. For consulting and services firms, this becomes a market differentiator: they can offer clients tested blueprints for agentic AI workforce transformation, backed by their own large-scale deployments. As AI agents take on more operational tasks, the ability to prove governance, audit trails and responsible AI practices will likely be as important as raw productivity gains in winning and retaining clients.
