Artificial intelligence (AI), especially machine learning, is rapidly finding a role across a variety of enterprise functions. One that you may not have considered is AI’s applicability to tax. Here are five ways that AI is reinventing the world of tax functions.
One of AI’s most simplistic—yet useful—applications involves the automation of routine tasks, the kind of time-consuming work that can fill a professional’s workday. This work involves using machine learning algorithms that have recently become common in today’s business world and our daily lives. For example, optical character recognition and voice recognition are being adapted to support the tax function.
These intelligent algorithms can be put to work to categorize sales into the proper jurisdiction for managing sales tax calculations. Or they can scan and analyze employee expense reports to determine which deductions are available based on the type and amount of each expense. While using machine-learning algorithms requires some level of thought, they are not overwhelmingly complex, and, in fact, some of these tools can be installed and managed by a typical business or tax professional.
Since the rise of the cloud computing era, the dashboard–which can provide suggested actions based on learning algorithms–has become one of the most indispensable business tools available. While a simple dashboard may give a tax professional a look at that day’s receipts and sales tax liabilities, a dashboard informed by AI is more dynamic, allowing a tax professional to use the system not only to evaluate the past, but also to anticipate the future.
Is the tax burden increasing beyond expectations because of an unexpected disparity in regional sales? Is it time to sell an investment to reap a tax benefit before rules or market conditions change? AI-enabled dashboards can give tax pros an instant understanding of these issues at a glance, aiding decision making.
AI will also make forecasting more accurate and faster, trading simplistic modeling techniques (such as basic regressions or linear interpolations) in favor of tactics that can detect trends on an annual, quarterly, monthly, or even more frequent basis. These trends can be drawn from corporate data or by looking at seasonal changes. A savvy AI algorithm may even be able to look at weather patterns to determine how sales are likely to be impacted regionally and the impact that could have on tax burdens.
What if you could calculate the probability of being audited based on potential sales and expense scenarios for the coming year? AI will make this type of analysis possible—even commonplace—by analyzing internal tax data and comparing that with available information about the broader economy.
Would moving the company headquarters to a new jurisdiction increase the chances of an audit? What about changing the business’s legal structure? AI might even be able to analyze what would happen to a business’s tax risks and overall structure based on the likely outcome of a local or national election, all by looking at a broad data set of unstructured data and building a probabilistic model out of it.
At the highest level of AI function, machine learning will be put to task to make (or aid with) complex decisions in the absence of any real structured data at all. For example, instead of your legal team spending hours poring over a 500-page sales contract, an AI tool will be able to do the same job in a matter of seconds, determining whether the document contains any thorny legal problems or taxation risks.
Similarly, an AI system may keep constant watch over a business’s register of sales actions, making educated decisions about whether the company is at risk of fraud. This kind of analysis, when performed by a human, often comes down to experience and intuition; a human fraud analyst simply “knows what to look for.” An AI tool can build its own statistical model based on historical fraud rates while looking for patterns that may be missed by a human examiner.
Artificial intelligence is just taking its first steps into the meticulous and complex world of taxation. As this trend accelerates, watch for AI to become an increasingly crucial part of progressive—and successful—tax departments.
Principal, Advanced Tax Analytics & Innovation, PwC US