How artificial intelligence is changing accounting

Artificial intelligence isn’t coming. It’s already here.

Early investments by large firms, including several of the Big Four, have paid off with advanced technology that can, among other things, slash the amount of time accountants spend on complex audits and asset estimates.

Now all firms, even small ones, should be thinking about how to adopt advanced technology like artificial intelligence (AI), whether it will be by contracting with specialized technology companies or building their own departments, said Derek Bang, CPA, CGMA, the chief strategy and innovation officer at Chicago-based Crowe Horwath, one of the largest accounting firms in the United States.

“They’re going to have to have a strategy,” Bang said. “You need to start putting some money aside into innovation.” 

Reports of machines replacing accountants by the thousands are likely overblown, and capable accountants will be needed to oversee and utilize advanced technology, said Tom Davenport, a Babson College professor who studies business applications of AI.  

“At some point, there might be some job loss on the margins, but mostly we’re giving people more sophisticated work to do than just looking through documents,” Davenport said.

AI explained

AI is technology that enables computers to perform decision-based tasks previously left to humans. It shows up in multiple forms, including machine-based learning that can progressively become better at analysis and decisions the more it is used, and speech-based technology that can understand different voices and languages.

More than 80% of executives believe AI leads to a competitive advantage, and 79% believe it will increase their company’s productivity, according to a recent MIT-Boston Consulting Group survey of more than 3,000 business executives.

AI is being developed by multiple accounting firms and will dramatically change the profession in coming years, said Jon Raphael, CPA, Deloitte’s audit chief innovation officer. 

“We’re at a real pivot point in terms of being able to wrangle and use data in ways we’ve never even contemplated before,” he said.

That doesn’t mean walking, talking robots like Star Wars‘ C-3PO and R2-D2 will be handling client meetings anytime soon.

Instead, AI is largely being used to digest and analyze large volumes of data at speeds well beyond what any person or team of people could do, Davenport said.

“What AI is doing is sort of a more sophisticated version of what spreadsheets do,” he said. “The more analytic and decision-oriented computations, at least for the next 20 to 30 years, will still require humans.”

Full speed ahead

At Deloitte, auditors can access AI tools with natural language processing capabilities to interpret thousands of contracts or deeds, Raphael said. The technology can extract key terms and compile and analyze that information to perform risk assessments or other functions.

Entire populations or datasets — a large company’s leases, for example — can be assessed in a shorter period of time, whereas the samples previously examined by accountants could have taken a lot longer.

“It’s creating a new view about where the profession is going,” Raphael said. 

Applications of AI and machine-based learning vary from firm to firm, with huge variances in how companies develop the technology, Bang said.

“AI and machine learning is as deep a field as accounting is,” he said.

That means firms can differentiate themselves and corner areas of the market by specializing in different areas.

At Crowe Horwath, Bang and his team of 20 data scientists have harnessed technology to tackle complex billing problems in the health care industry. The team was able to use machine-based learning to sift through enormous but disparate billing systems of its health care clients to flag accounts where there are time-consuming and costly complexities in claim processing and reimbursement. In those cases, the health care companies and hospitals can use the technology developed by Crowe Horwath to proactively deal with those complicated cases instead of waiting for the problems to make themselves known, a development that can save clients hundreds of man-hours.

“We’re finding we can solve problems that couldn’t have been solved otherwise,” Bang said.

Instead of talking about hypothetical future uses, Bang prefers to show his firm’s leaders what he and his team have already developed, from health care billing solutions to conducting the background research needed to apply for research and development tax credits.

At Deloitte, the excitement about AI is shared by its clients, with auditors showing how they are using AI-based applications to quickly conduct accurate assessments of vast real estate holdings or analyze thousands of contracts to compile the risks a large company faces.

“Clients want to do this in their own business,” Raphael said. “We spark the thinking of where they can go.”

How to prepare

To prepare for the oncoming wave of AI, Raphael suggested that interested CPAs gain database and IT skills by taking on specialized projects in their workplace, attending seminars, completing self-directed learning, or taking classes.

Having a solid basis in data management and a high comfort level with new technologies will give those practitioners an edge as AI use increases in the field.

He said that the professional skepticism auditors are trained to have is needed as well to be able to spot when the analyses are off and to deal with exceptions.

“We don’t want to have people relying on a tool blindly,” he said.

What’s ahead

Smaller firms don’t have the same resources as largest firms to develop and fine-tune their own AI products. But experts say that the technology will be more widely available in coming years, and they expect it eventually to become standard fare.

Davenport likens it to the advent of the internet. When the World Wide Web was first publicly available, only a few large companies could afford to go online or develop their own intranets. That changed, of course, and today there is scarcely a firm, or person, left unconnected to the web.

The same will happen with AI, and it may become as common as the internet is now, Bang said.

“The accountants of the future will exist, but they will know how to interact with machines,” he said.

Sarah Ovaska-Few is a freelance writer based in North Carolina. To comment on this story, contact Chris Baysden, senior manager of newsletters at the AICPA.