Big Four Firms Roll Out AI That Can Handle Routine Tasks Solo

The next generation of artificial intelligence has arrived at the Big Four firms, promising to upend the core work of their legions of accountants and business consultants.

Deloitte, the largest of the four, and competitor Ernst & Young said this week that they are diving into new autonomous AI tools that can walk clients through routine tasks like uploading documents needed for tax returns and take on more complex work analyzing financial statements. The technology falls into a category known as agentic AI.

PwC, also known as PricewaterhouseCoopers, has also developed such capabilities, known as agents, to speed up software development and tailor guidance for auditors. KPMG is not far behind its larger peers and is developing “digital teammates” to support groups across its global business.

The moves mark a new milestone in the four firms’ technology race. Combined, the Big Four have pledged billions of dollars to bake AI into core platforms and guide their blue-chip clients to make the most of the fast-developing technology.

“You can’t sit back on your heels and wait to figure out the technology,” said Nicole Wright, an associate accounting professor at James Madison University. “You have to get ahead of the curve and understand how it’s going to fit and how you’re going to provide it to clients.”

Both EY and Deloitte have partnered with Nvidia Corp., combining the chip maker’s AI architecture with their in-house data and technical expertise in tax law and accounting standards. The goal is to speed up common but manual and sometimes time-consuming tasks like researching how foreign tax rules apply to a client.

Their launch announcements earlier this week coincided with Nvidia’s annual developer conference, dubbed the “Super Bowl of AI.”

The latest tools based on agentic AI are intended to free up employees for more complex work rather than replace them, but could have a dampening effect on future staff needs.

AI agents will deliver an “immediate” productivity boost to EY teams, Raj Sharma, global managing partner for growth and innovation at EY, said in an interview. “When they are going to the clients, now they are equipped with 100 years of knowledge in that particular domain at their fingertips.”

The New AI

Agentic AI builds on the content creation abilities of its predecessor known as generative AI, which emerged as the technology behind ChatGPT and other platforms. Content generator models can provide a list of ideas useful for research or drafting emails to clients.

But this latest iteration of AI performs tasks with little human intervention, said Subodha Kumar, a business data science professor at Temple University.

“Agentic AI as a model is designed to solve the problem for you,” Kumar said. “It is based on the end goal.”

Armed with $1.4 billion, EY plans to roll out as many as 150 different agents in the next three months to 80,000 tax professionals around the globe, Sharma said.

EY chose to deliver the new autonomous AI tools to its tax practice first in part because of its extensive well of in-house tax data and the varying needs of clients doing business globally, he said.

Deloitte’s early agents are designed for use by finance teams, including the firm’s own staff who are employing it to manage business expenses. Its partner Hewlett Packard Enterprise Co. has begun to use Deloitte agents for financial statement analysis and other tasks like modeling financial scenarios, according to a statement.

The firm committed $3 billion last year to building out AI solutions and relationships with heavyweights like Google and Nvidia.

Deloitte aims to support its clients “in their transition into this new era, where agents and employees interact to reinvent business processes and unlock new sources of business value,” Jason Girzadas, CEO of the firm’s US arm, said in a statement.

KPMG said it is building agents both for use by its professionals and its clients with a focus on customer service and quality, among other areas.

Meanwhile, PwC has teamed up with Microsoft Corp. and Salesforce Inc. to build agents for its clients and announced a similar partnership with Oracle Corp. on Thursday.

“Agents can help unlock the value of AI for the enterprise. We’re entering a period of agentic organizations that will fundamentally change how we work across functions and industries,” Dan Priest, chief AI officer for PWC US, said in a statement.

Staffing Needs

Agentic AI technology is meant to supplement, not replace human employees. But the digital agents are taking on work currently shouldered by thousands of Big Four employees and may mean firms won’t need to expand their payrolls as quickly in the future.

At KPMG, staff will be paired with what David Rowlands, the firm’s global head of AI, called a workforce “that ignores siloes, borders, politics and never sleeps.”

Deloitte said in a statement that its finance agents would “liberate thousands of hours” per year, slashing costs by as much as 25%.

EY staff will be able to serve a larger roster of clients than they do today with the help of agentic AI. As the firm expands, that AI-powered workforce won’t need to grow as quickly, Sharma, the global managing partner at EY, said.

All four firms’ AI commitments include workforce training to prepare teams not only to use AI but also to stay ahead of the fast-paced tech and the needs of their clients.

Despite the promise to increase productivity, the models risk returning the same answers repeatedly, stunting creativity and the profession’s ability to adapt to changing regulations and markets developments, Wright said. But she added she is optimistic that skilled accountants will remain in demand.

“You still need that human person on the other end who knows the background, who knows the actual laws and the requirements and they can verify that yes, what it is saying is appropriate, is accurate,” Wright said.