If you are looking for proof that "AI agents have entered production environments," EY is currently the most compelling example.
In early April, EY announced the global deployment of enterprise-scale agentic AI, covering its 130,000 audit professionals, spanning 160,000 audit engagements, across 150 countries and regions.
This is not a pilot. This is a production-grade global rollout.
How big is the scale
EY used one metric to measure the scale of its system: processing 1.4 trillion rows of journal entry data annually.
What does this number mean? Simply put, any manual process is infeasible at this scale. Just running through the data already exceeds the boundaries of human capability — this is the real reason AI agents exist, not to "improve efficiency by 20%," but because "no human can do it at all."
EY's core platform is called EY Canvas, a unified technology platform covering all audit engagements globally. The new agent system is embedded directly into it, built on the underlying capabilities of Microsoft Azure, Microsoft Foundry, and Microsoft Fabric.
EY is one of the first members of Microsoft's Frontier Firm AI Initiative, with only 14 organizations globally recognized for deploying advanced AI at this scale.
What the agents actually do
According to information disclosed by EY, this multi-agent system primarily handles three types of work:
- Coordinating complex audit tasks: Orchestrating tasks across multiple processes and technology systems without requiring auditors to manually connect them
- Dynamic risk processing: Updating risk assessments in real-time based on new information, providing continuously updated audit and accounting guidance
- Reducing administrative burden: Delegating repetitive, low-judgment tasks to agents, allowing auditors to focus on areas requiring professional judgment
EY specifically emphasizes: AI handles the workload, but human judgment, professional skepticism, and insight remain at the core.
This statement is not just PR rhetoric; it is also a regulatory requirement for the audit profession — registered auditors are responsible for the final opinion, not AI.
How much did it cost
EY describes it as "multibillion-dollar" — meaning in the tens of billions of dollars range.
The exact amount has not been disclosed, but the magnitude itself speaks volumes: this is not a small project within an IT budget; it is a strategic investment to redefine how the audit industry operates.
The full deployment timeline extends to 2028. The current phase is already running, with complete end-to-end AI-supported audits expected to cover all engagements by 2028.
EY Global Chair and CEO Janet Truncale put it directly:
"EY's human-machine collaboration, AI-powered audit, is a market-leading example of enterprise AI in practice."
Why the Big Four are rushing
The audit industry faces a subtle pressure: clients are using AI, so auditors must also use AI to audit AI-generated outputs.
An increasing portion of a listed company's financial statements is AI-generated or AI-optimized. Reviewing these numbers manually would lead to issues with both efficiency and accuracy.
EY's logic is: it's not that we want to embrace AI; it's that the industry has reached a stage where audits cannot be done well without AI.
More immediate competitive pressure comes from comparisons among the Big Four. KPMG, Deloitte, and PwC are all advancing their own AI deployments simultaneously. Whoever establishes a first-mover advantage in scaled deployment will be able to win more clients in the coming years.
What it means for auditors
EY has simultaneously launched a training program covering all audit and technology risk personnel globally, and has joined Stanford University's Institute for Human-Centered AI (HAI) Industrial Affiliates Program.
Judging by these supporting actions, EY does not view this as a simple "tool upgrade," but as a transformation requiring the retraining of an entire professional workforce.
The other side of this deployment is also worth noting: if AI agents can handle a significant amount of audit work, how many people will a team of 130,000 need in the future?
EY's answer is 2028, when the situation may become clearer.
The accounting firm industry has always been known for its conservatism. EY's move shows one thing: even the most conservative industries are starting to take agents seriously.
Sources: EY launches enterprise-scale agentic AI to redefine the audit experience for the AI era (EY official press release); EY Rolls Out Agentic AI in Assurance Across Its Global Network of Accounting Firms (CPA Practice Advisor); CocoLoop; EY's agentic AI pivot - A watershed moment for audit quality? (Accountancy Age)