EY announced on April 7, 2026 that it has embedded a multi-agent AI framework directly into EY Canvas — its single global audit platform — making this one of the largest real-world deployments of agentic AI in professional services to date.

The rollout is not a proof of concept. It covers 130,000 Assurance professionals working across 160,000 audit engagements in more than 150 countries. The platform already processes 1.4 trillion lines of journal entry data every year. Now, AI agents are part of that workflow.

What EY Actually Built

The initial deployment consists of one core assistant agent and three supporting agents, collectively handling roughly 20 modular capabilities. The early focus: searching and summarising documentation, and automating administrative tasks that currently eat auditor hours without adding much professional value.

Two more agents are already in development. One will review auditors’ work papers and suggest improvements. Another will focus on reconciliation documentation. By 2028, EY expects agentic AI to support all end-to-end audit activities across the firm.

The technical stack is built on Microsoft Azure, Microsoft Foundry, and Microsoft Fabric. EY is also one of only 14 organisations selected for the inaugural class of Microsoft and Harvard’s Frontier Firm AI Initiative, a joint programme studying how frontier firms integrate AI into professional workflows.

The whole push sits within EY’s broader “All in” global strategy and a multibillion-dollar commitment to audit quality, technology, and people.

Why This Matters More Than a Product Launch

Most enterprise AI deployments at this scale are announced before they are real. This one is already running. And the scale makes it worth paying attention to: 130,000 professionals, 160 countries, the core audit workflow of one of the four largest accounting firms on the planet.

This is how adoption actually moves. Not a single company experimenting with one agent in one department, but a coordinated rollout into a critical, high-stakes knowledge process with regulatory consequences. Auditors cannot afford hallucinations or unreliable outputs. The fact that EY committed to this at scale says something about where enterprise confidence in agentic AI now sits.

For the professional services industry broadly, the signal is clear: the Big Four are moving. Deloitte has been building agentic workforce capabilities through Zora AI. EY is now running agents on 1.4 trillion lines of financial data. This is not a future trend. It is a current one.

What This Means for Business

For leaders in professional services: If the largest audit firms on the planet are building AI agents into their core delivery workflows, the question for your firm is not whether to do this eventually. It is how far behind you are willing to fall in the meantime.

For data and finance teams: The EY deployment shows what AI agents are good at in this context: synthesising large volumes of structured data, summarising documentation, automating repeatable administrative processes, and flagging inconsistencies for human review. These are tasks most finance teams also deal with. The technology is proven.

For technology leaders: The architecture decision here is notable. EY is not building on a single model vendor. They are using Azure infrastructure, Microsoft Foundry for model access, and Microsoft Fabric for the data layer. Multi-agent frameworks sitting on top of enterprise data platforms — this is the pattern that is emerging across serious deployments.

For anyone still in pilot mode: EY moved from experimentation to global deployment. The pilot-to-production gap is closing faster than most organisations expected. The window to start learning is now. The window to have caught up is already passing.

---

The argument for agentic AI in complex knowledge work has often been theoretical. EY just made it operational, at a scale that leaves little room for “we’re watching to see if it works.” It works. The question is what your organisation is doing about it.

For a deeper walkthrough of tools like this and how they fit together, the free Working With Claude field guide covers the ecosystem end to end. Get the guide.