KPMG Rolls Out Generative AI to Tax Pros, Launches Audit Pilot

Inside KPMG, its tax professionals are leaning on generative artificial intelligence to help corporate tax departments prepare for new requirements to disclose their tax obligations by country. Its clients have been given access to that “virtual assistant” for help gathering tax data and guiding them through the analysis and drafting of reports on what taxes were owed around the globe.

Those are just a few ways that KPMG’s $2 billion commitment already is integrating the next level AI into everyday business. The firm, leveraging its more than decade-long partnership with Microsoft, is tapping the tech giant’s Azure cloud platform and OpenAI, the creator of ChatGPT, to help its accountants and consultants provide faster insights.

KPMG has said it would co-develop cloud and AI-based products for its employees and its clients—a five-year investment that the firm projects could return as much as $12 billion in revenue.

Combined, the Big Four firms are funneling billions into the innovative technology, building on past investments to reinvent their audits and fuel their acquisitions.

“This is probably the most transformative moment for technology in a generation,” said Brad Brown, global chief technology officer for KPMG’s tax business of the expanded Microsoft partnership and AI’s impact.

Productivity Boost

KPMG’s entire tax practice now has access to ChatGPT, and the firm is working to add it to products on which its clients also rely, Brown said in an interview with Bloomberg Tax. “In essence, our professionals have got a new team member.”

That tool sits within KPMG’s firewall, and the firm dictates which data the tool can access within those protections. Staff mustbe trained to gain access to the tool.

Past investments laid the foundation for KPMG to embed this latest technology into its products and across service lines, Brown said.

KPMG auditors may eventually be able to use the technology to similarly analyze disparate data. For now, the firm is running pilot tests focused on tasks such as summarizing contracts or preparing presentations for audit committees, said Thomas Mackenzie, global chief technology officer for KPMG’s audit practice.

The focus is on boosting productivity, Mackenzie said, describing the firm’s plans for using generative AI for audits as “cautious” and “careful.”

“We need to ensure that the results that are leveraged on an audit are certified, verified, understandable, and can be used to drive audit quality,” he said. “That is why we are taking a very measured approach around adopting AI because we cannot have someone rely—nor can tax or advisory for that matter—rely on results which are malformed due to a misunderstanding or misapplication of AI.”

Big Four Accelerate AI

Like KPMG, the rest of the Big Four have launched Gen AI pilot programs and services to get ahead of the curve and make sure their workforce and services are ready when their clients are.

The investments feed into the billions the firms have collectively spent to upgrade their tax and audit suites in recent years and to train thousands of accountants to use the new tools to wrangle financial data, spot anomalies, and identify risks.

Deloitte, which is limited from partnership deals with its audit client Microsoft, has instead teamed up with other tech giants such as Amazon Web Services, Google, and NVIDIA to provide cloud and AI-based offerings. The firm also has bolstered its team and AI offerings through a series of acquisitions including HashedIn Technologies and Intellify, Deloitte said.

In April, the firm launched a dedicated generative AI practice to help clients develop their own strategies for the evolving technology and in December it committed $1.4 billion to train its staff in tech areas such as AI.

PwC in April announced a $1 billion investment to expand how its US professionals use AI and to train those 75,000 workers to use the technology. That training will cover the basics of generative AI including how to write queries, the firm said.

The firm also is testing the use of artificial intelligence to help its auditors with routine tasks such as pulling out key terms and comparing data, or summarizing meeting minutes and contracts, PwC said.

Ernst & Young LLP in June said it would train its global tax workforce on how to use the same Open AI tool as KPMG. The firm also has a number of partnerships with ServiceNow, Adobe, and NVIDIA, and it will provide Microsoft 365 co-pilot, another AI-based digital assistant, to boost productivity. An advisory council will guide its use of AI, the firm said.

Beyond the Big Four, firms of all sizes must gear up for the emerging tech by using it for simple tasks such as drafting emails that don’t rely on client data or confidential information, said Amy Vetter, CEO of the B3 Method Institute, a training and consulting group that helps accountants implement technology.

In time, the tool will surface information for accountants faster, allowing them to quickly pull pertinent information from records and contracts. The technology won’t eliminate accounting jobs but should transform how accountants work,elevating their expertise, Vetter said.

“No one in accounting wants to work the amount of hours that they work,” Vetter said. “So if there’s technology that can help that, that doesn’t devalue the accounting profession.”

Risks and Rewards

For now, firms are jockeying for a share of that burgeoning estimated $96 billion global AI market—posturing to future clients who may soon be ready to add Gen AI to its team or corporate directors eager for a more efficient, effective audit.

Still the technology’s potential to eclipse the capacity of human intelligence poses a risk to firms that have historically sold the knowledge and expertise of their staff, said Isaac Heller, CEO of accounting tech company Trullion.

“It’s a gut punch in the stomach to people who’ve built their careers on their knowledge accumulation,” Heller said. “It’s a true existential threat. It doesn’t mean they can’t turn that risk into an opportunity.”

Buying into the hype that generative AI is an all-purpose tool is also a risk. The tool may not benefit from the same level of training in certain countries or functions—leaving potential gaps for large international organizations such as the Big Four accounting firms, said Masheika Allgood, founder of AllAI Consulting LLC and an AI ethicist.

“Treat it the same way you treated the other AI you rolled out in your company, which was for specific purposes,” Allgood said. “We’re taking the complete opposite approach to generative AI, and I think it’s going to cost some businesses a lot of money and reputational damage.”

Business leaders are eager to know more about its potential but companies aren’t rushing in to deploy generative AI just yet, said Kevin Rubin, the CFO of Alteryx, an enterprise analytics platform used by large companies around the world.

“Our customers are absolutely interested in understanding how Gen AI can improve their businesses,” Rubin said. “But they’re doing so in a cautious, responsible way and wanting to be able to understand all of the implications of deploying these technologies in their businesses before jumping in.”