4 Ways AI will Revolutionize the Tax Function
What if organizations could automate tax classification work that currently takes thousands of man-hours to complete? Or process millions of transaction records within seconds? This may seem as a far-off fantasy, but AI is set to revolutionize the tax function of the future.
Tax professionals have traditionally been the guardians of compliance in a complex field. As innovative technologies come to fore, their roles are beginning to change. At a time when information is more accessible than ever, there is a massive opportunity for tax professionals to move from being a responder to being a crucial part of developing and implementing strategic business change.
Although tax professionals are more accustomed to disruption than their other organizational counterparts, as legislation is always changing. There has also been a marked shift in recent years towards risk and compliance management in light of the more transparent and data-driven organizational culture.
Cognitive computing and machine learning have already digitized and automated routine tax compliance to a certain extent. AI capabilities already in use for OCR recognition (ATM checks), and voice recognition could be readily adapted to other arenas of tax compliance as well. At higher levels of tax functions, AI enabled tax applications may address more complex, human-judgment tasks like answering subtle legal and taxation questions from a document or detecting fraud, thereby also assisting government oversight.
Four ways AI will disrupt the tax function:
- Forecasting: This is perhaps the simplest and the most obvious application of AI and predictive analytics. To start with, an algorithm first determines a trendline and based on that it can either extrapolate or interpolate the trend to make a prediction using new data.
Although most tax professionals use basic techniques for tax provision forecasting, **AI can be a gamechanger in terms of detecting trends within various tax filing cycles (annual, quarterly, etc.)**. In other words, these algorithms address a simple question: “if current trends continue, what will happen next?”
- Clustering: From the time of early stargazers, humans have mastered the art of pattern recognition. Our minds have an amazing ability to recognize different types of patterns and then transform these into concrete actionable steps. Maybe this is why people with great pattern recognition skills make exceptional chess players. However, as the famous Garry Kasparov vs IBM Deep Blue match demonstrated, computers are simply faster and more accurate at identifying these patterns. What’s more? They can scale!
Clustering is particularly useful for recognizing specific business behavior, such as:
- Companies that have similar tax characteristics
- Expense generating behavior of employees or businesses
- Grouping of tax implications in transactional data including, sales tax, capital gain, indirect taxes, and detection of unusual behavior in expenses and purchasing.
Clustering is the primary technique used in tax data analytics. This form of “unsupervised” machine learning is exploratory in nature and often reveals insightful behavioral traits of individuals and companies.
- Virtual Support Systems: **AI-powered virtual support systems (like Siri, Alexa and Google Home) can be used to assist tax professionals by suggesting answers to detailed questions and complex problems**. AI-powered platforms like IBM’s Watson, that can be trained to answer specific questions in a particular field, like finance, sports or general trivia are being actively adapted to various organizational functions.
- Natural Language Processing: At the confluence of AI and linguistics, Natural Language Processing (NLP), helps automate the identification of “meaning” in text – in other words, NLP is capable of interpreting human language. Today, NLP is being widely used in conversational interfaces and language tools. In the realm of tax compliance, NLP systems can imitate human learning patterns to process vast amounts of textual data and refine identified meaning to offer actionable insights to tax professionals.
The new tax ecosystem
The tax function today is in a state of flux and even bigger changes lie ahead. It’s impossible to predict where we’ll be in a decade, or even in five years’ time. But what’s clear is that artificial intelligence and machine learning are catalysts and enablers of this new tax ecosystem. The demand for new capabilities and roles in the tax function will dramatically alter its relationship with the business. The tax function is ripe for disruption, and so it is critical for tax professionals to understand what might lie ahead, and AI is just the beginning.
