AI Poised to Revolutionize Tax Preparation

Advances in ChatGPT and consumer Generative Artificial Intelligence (AI) interfaces have propelled the technology into the spotlight, inspiring breathless media accounts of dangerous implications for civilization and Luddite outcry over its job-killing potential. While the science fiction hype may be overblown, AI’s potential world impact is massive.

Generative AI is compelling because the technology understands and interprets basic language and crafts unique, humanlike responses, unlike earlier machine-learning technology called Analytical AI.

In a note issued on March 26, Goldman Sachs Chief Economist Jan Hatzius wrote, “The boost to global labor productivity could also be economically significant, and we estimate that AI could eventually increase annual global GDP by 7 percent.”

“Although the impact of AI will ultimately depend on its capability and adoption timeline,” Hatzius added, “this estimate highlights the enormous economic potential of Generative AI if it delivers on its promise.”

According to Goldman’s research, AI could drive upwards of $7 trillion in global economic growth over the coming decade.

AI promoters argue that Generative AI, when combined with more analytical proceeding technologies, will, among other things, allow the average person to navigate complex data sets without special knowledge.

No task would appear more suitable as an application for AI than tax preparation.

Maddening Complexity

“Today, it takes more brains and effort to make out the income-tax form than it does to make the income,” reads a quote attributed to Mad magazine’s fictitious Alfred E. Neuman, often repeated during discussions of US tax policy.

“Major, structural tax reform does not happen often and tends to be a once-in-a-generation event,” the Tax Foundation stated in a 2021 report.

The report from the center-right think tank founded by corporate leaders during the Great Depression adds the understatement, “[and] between major reforms, the tax code tends to get more complex, not less.”

Congress’s approach to taxing individuals and businesses in the US has become increasingly complex over the past century, as it employs the laws governing taxpayers as a mechanism for patronage and advancing policy goals.

The actual length of the US Federal tax code is a matter of ideological dispute, as the numbers represented by lawmakers and activists indicate their slant. Title 26, the laws enforced by the Internal Revenue Service (IRS), features 2,600 pages of regulations. IRS regulations, revenue rulings, and other clarifications add thousands of supplemental pages to the tax code to guide filers.

Politicians and pundits arguing for tax policy simplification — typically conservatives — include this case law in the count to arrive at a more odious sounding 70,000 pages.

Temporary provisions in the tax code demand an annual review to ensure compliance. State tax laws, which can be equally dense, add another layer of due diligence. As a result of the massive amount of regulations to navigate, the business of tax preparation — particularly business and complex personal filings — is incredibly lucrative.

According to projections by research firm IBISWorld, the US tax preparation industry is a roughly $14.4 billion market, measured by revenues in 2023. As the tax code has increased in complexity, the industry has expanded faster than the US economy overall, with an average annual growth rate of 3.9 percent between 2018 and 2023.

Poised for a Breakthrough

A 2020 survey from PC Magazine reported that 58 percent of respondents who prepare their taxes use the TurboTax platform from Intuit. Currently, TurboTax boasts 100 million individual and small business customers.

Intuit already utilizes AI technology in its Virtual Expert Platform, making it one of the first firms in the space to adopt the technology commercially. The language processing in TurboTax allows users to use chat to navigate returns, although the Virtual Expert Platform still refers filers with complex needs to human experts.

Industry research suggests that taxpayers with the most complicated filings, including family offices, large corporations, and companies with international business operations, still rely on accounting and law firms. While software provides powerful tools for professional preparers, it lacks the ability to decide how aggressive to be in interpreting laws or anticipating a client’s future needs. It also requires constant upgrades to stay abreast of code changes and interpretations.

AI offers the potential for machine preparation capable of staying up to date on changing regulations independently and interpreting both client risk tolerance and current case law to make judgment calls.

“I predict that the tax software will get smart enough that it will assist the professionals in taking most advantage of the tax code,” says Randy Johnston, a shareholder in K2 Enterprises. “The hardest part will be the risk tolerance because some CPA professionals take a lot of risk with their clients while others are very conservative and take very little risk,” Johnston adds. “A lot of that stuff is opinion, but these tools will give an opinion.”

Johnston has over 40 years of experience in computer science and worked on early AI research during the 1970s. Today, his work at KZ focuses on financial technology with a focus on accounting. According to Johnston, K2 has worked with a majority of the top 100 CPA firms.

A Changing Competitive Landscape

While large tax software providers like Intuit and H&R Block have plowed tremendous resources into proprietary AI technology, some industry analysts see the consumer technology leveling the playing field.

In a note released on March 16, a team of analysts at RBC Capital Markets led by Rishi Jaluria concluded that disruptive midmarket software firms could harness Generative AI to go after the entrenched market share of legacy competitors.

Traditional human tax experts who manage complicated client filings manually, including elite accountancy, will also potentially feel pressure.

“I think there’s a heavy threat to white-collar workers for their knowledge,” says K2’s Johnston. “Now, that doesn’t mean they’re going to be eliminated. What will happen is the very best of the workers will use these tools effectively to expand their practice.”