BETA
This is a BETA experience. You may opt-out by clicking here

More From Forbes

Edit Story

EY, Deloitte And PwC Embrace Artificial Intelligence For Tax And Accounting

This article is more than 6 years old.

Tax preparation, auditing and strategy consulting are services that have historically relied on an intensive use of human capital. Artificial intelligence (AI) now threatens these business models. Technologies such as natural language processing (NLP) and robotic process automation (RPA) complete in mere hours what human auditors took weeks to do. Deloitte, Ernst & Young (EY), and PricewaterhouseCoopers (PwC) are three of the famed Big Four tax and accounting companies. I recently sat down with innovation leaders at these companies to learn how AI is transforming the tax and auditing business.

Approaches to AI

Each of the three companies, Deloitte, EY and PwC, employs a slightly different process for developing AI technologies.

EY starts small and aims to demonstrate immediate ROI. Chris Mazzei, chief analytics officer and emerging technology leader at EY, explains, “You have to take it from a business value perspective first, rather than a tech perspective first.”

PwC, especially in its client services business, favors the use of four-week-long AI “sprints” that mirror the Agile software development method. The goal is to quickly demonstrate a working model for a client before refining it to a higher accuracy. “Literally in a month, we take something from a concept to a first implementation,” said Anand Rao, global artificial intelligence lead at PwC. These sprints first started three and a half years ago. Today, the company holds 70-80 sprints each year. Each sprint is staffed by a small team of just two to three people.

At Deloitte, a 70-member internal corporate innovation team focuses on all aspects of emerging technology; the team devotes 80% of its time to AI. Its mandate is to create use cases that guide AI-related investments across Deloitte and also for external clients. Craig Muraskin, managing director of innovation at Deloitte, described the role of the team as being “an accelerator of innovation activities across each of our businesses.”

Here are some successful cases of AI use in tax and accounting:

Natural Language Processing (NLP) for Document Review

One of the best examples of AI put to use in accounting is in the review of high volumes of contracts. According to Muraskin, a Deloitte team might comb through hundreds of thousands of legal documents looking for a change of control provisions during a client’s sale of a business unit. This used to keep dozens of employees occupied for half a year. Now, a team of six to eight members can use an AI system to complete the same task in less than a month.

Similarly, EY uses AI to review lease accounting standards. When the IRS issues a new lease regulation, large companies have to manually re-examine tens of thousands of leases to comply with the new law. Using NLP to extract information and a human-in-the-loop to validate the results, the AI system is three times more consistent and twice as efficient as previous humans-only teams. This system reached break-even ROI in less than a year. Nigel Duffy, global innovation artificial intelligence leader at EY, noted that the company not only benchmarked AI/ML processes, but also the human processes. Now [they] have better measurability with quality and costs.”

Evaluating potential procurement synergies during mergers and acquisitions requires looking at hundreds of millions of lines of non-standardized accounts payable and receivable data. Reconciling all this data is a nightmare that used to involve manually building spreadsheets and pivot tables. And it took so much time and effort that no one dared to do it until after a deal with completed. Today, intelligent classification engines quickly compile the information in AI systems that learn. At Deloitte, what once took four to five months to complete is now done in a week , which means the analysis can be done during deal evaluation. Muraskin said, “It opens up all sorts of opportunities to bring greater value to the client. It opens up new business opportunities.”

Machine Learning for Anomaly Detection

Anomaly detection, which is the identification of outliers in data, is a classic application of machine learning. Similar to how a credit card company identifies potentially fraudulent credit card charges, EY uses ML to identify fraudulent invoices for clients. One of EY’s global clients processes millions of invoices each year and uses this system to identify the faulty ones. This service is especially valuable for detecting fraudulent invoices that involve international parties and, in turn, helps to avoid the serious consequences that result from violating sanctions, anti-bribery regulations, or other aspects of the Foreign Corrupt Practices Act. EY’s fraud detection system has a 97% accuracy and has been rolled out to over 50 companies. Mazzei shared that “rather than using project-based calculations for time and materials, this is a new way of doing business. This can be a new business model.”

Natural Language Generation for Producing Reports

Deloitte uses natural language generation (NLG), the creation of text by computers, in its tax business. The company processes upwards of 50,000 tax returns annually for clients’ employees who have expat status or other complicated financial situations. Using NLG, Deloitte creates detailed narrative reports of individual tax returns. Its tax professionals rely on these reports to provide more targeted financial advice to clients during consultations.

Innovation leaders at the Big 4 are actively deploying AI. As Steve Toy, innovation lead at EY, warned, “If you don’t innovate, the kid in the garage down the street will. Her technology may not be there today, but it will get better over time and soon replace your human-based processes. You need to innovate, or you will become obsolete.”

Moving forward, the ways in which AI will continue to impact businesses is perhaps best summed up by Booz Allen Hamilton Vice President Angela Zutavern in her book The Mathematical Corporation, “The most powerful weapon in business today is the alliance between the mathematical smarts of machines and the imaginative human intellect of great leaders. Together they make the mathematical corporation, the business model of the future.”

Follow me on TwitterCheck out my website