If you’ve been waiting for proof that AI agents are ready for serious business use, the Big Four just provided it. Deloitte, EY, KPMG, and PwC, the four largest professional services firms on earth, are each running live AI agent programs at scale. Not pilots. Not experiments. Production deployments handling real client work.

Here’s what’s actually happening.

Deloitte: Zora AI

Deloitte’s entry is called Zora AI, and it’s one of the most detailed public examples of an agentic AI platform built for enterprise operations. Built on NVIDIA’s AI infrastructure including the Llama Nemotron reasoning models, Zora AI deploys autonomous agents across finance, human capital, supply chain, procurement, sales, marketing, and customer service.

The targets are concrete: Deloitte is aiming for a 25% reduction in costs and a 40% productivity increase through the platform. Their finance team is already using it internally for expense management automation, and they’re planning to roll access out to thousands of employees by the end of the year. Hewlett Packard Enterprise is an early client, using Zora AI for financial statement analysis and scenario modeling.

What makes Zora AI notable isn’t just the breadth of coverage. It’s that Deloitte built it to function alongside human workers, not replace them wholesale. The agents handle the time-consuming, repeatable layers of complex professional work so that the humans can focus on judgment, client relationships, and the parts that actually require expertise.

EY: 150 Tax Agents for 80,000 People

EY’s deployment is built around their tax practice. They’ve embedded 150 specialized AI agents into the EY.ai platform, and those agents are now supporting 80,000 professionals worldwide. The scope is significant: EY expects to process more than 3 million tax compliance cases and transform over 30 million tax processes within the next year.

The business model implications are interesting too. EY’s leadership has been direct about the fact that the traditional hours-based billing model doesn’t survive agentic AI. Their framework for the future is what they’re calling “service as software,” where clients pay based on outcomes rather than the number of people working on a problem.

That’s a fundamental shift in how professional services gets priced and delivered, and it’s happening now, not in some theoretical future.

KPMG and PwC: Orchestration at Scale

KPMG’s Workbench takes a different angle, focusing on multi-agent coordination that mirrors how human audit teams actually operate. Multiple agents collaborate on different parts of an engagement, with oversight structures designed to match how the work is governed. KPMG’s global head of AI described agents as “innovative digital team-mates.”

PwC has gone furthest on raw numbers, deploying 25,000 agents across client operations through their Agent OS platform, built in partnership with Salesforce, CrewAI, and AWS. Their public framing for 2026 is telling: in 2025, AI was integrated into existing workflows. In 2026, the focus shifts to designing processes from the ground up with AI in mind. That’s a fundamentally different ambition.

The Structural Signal: Job Titles Are Going

Perhaps the most concrete signal of where this is heading came in January 2026, when Deloitte announced it would scrap traditional job titles effective June 1. The pyramid model that has defined consulting for decades, with junior consultants handling high-volume tasks and seniors overseeing them, is under direct pressure from agents that can handle the work that used to sit at the bottom of that pyramid.

KPMG’s language for this is worth noting. They’re preparing junior consultants to become “managers of agents.” The career path isn’t disappearing. It’s changing shape. The people who thrive will be the ones who know how to direct, evaluate, and improve the work that agents produce.

What This Means for Business

The Big Four are not early adopters. They are famously conservative institutions with enormous client liability exposure. When all four of them are running live production AI agent programs simultaneously, covering the most scrutinized parts of business operations, audit and tax, the question of whether AI agents are enterprise-ready has been answered.

The question now is whether your business is building the capability to benefit from this, or waiting while competitors build operational advantages that compound over time.

A few things worth considering:

The competitive pressure isn’t coming from startups. When EY automates 3 million tax compliance cases, they can take on more work with the same headcount, or the same work with lower costs. That pressure flows through to every business that competes for professional talent or services.

The productivity ceiling is moving. The 40% productivity targets Deloitte is running toward aren’t incremental. They represent a structural change in what a team of a given size can accomplish. Businesses that build equivalent capability in their own operations gain the same compounding advantage.

The governance model matters. Every one of these platforms includes human oversight layers, approval workflows, and audit trails. The Big Four didn’t deploy agents and hope for the best. They built the oversight infrastructure. Any serious AI agent deployment should do the same.

The practical next step is the free Working With Claude field guide. Thirty-two pages covering the ecosystem, Claude Code, and how to govern a rollout properly. Get your copy.