Eric Lefebvre, Chief Technology Officer at Sovos
For organisations trying to manage the frenetic pace of a rapidly evolving regulatory landscape, AI holds the promise of being a powerful tool across all areas of the business. Industries that deal with tremendous amounts of data, like tax compliance, are already starting to see the potential in machine learning and intelligent automation. Generative AI is a progression in technology that will allow everyone from the CFO to frontline compliance managers to solve use cases faster with off-the-shelf tools. It can work rapidly to make internal processes more efficient. That’s incredibly powerful, especially for smaller organizations, yet AI adoption – and its impacts – will be more evolutionary than revolutionary.
Business leaders should remember that we are at the beginning point when it comes to using this technology in business. There are tremendous benefits, but there are also real risks, such as the potential leak of intellectual property used in AI training. This amount of computing power has never been easy to implement, and it’s also easy to get wrong. Leaders undertaking an AI implementation should keep the adage in mind that with great power comes great responsibility.
With the right approach, there many use cases where AI can make a significant difference in compliance management.
Transforming Tax Tech & Regulatory Compliance
Tax compliance and regulatory reporting are ripe with opportunities for automation and AI implementation. For instance, in indirect tax reporting – such as for value added tax (VAT) and good and services tax (GST) – processes remain the same every month. AI can help organizations keep track of their tax obligations, flagging discrepancies as they come up, such as ensuring the right VAT is applied to a given transaction.
There are many similar process efficiencies. In purchasing, AI can facilitate decision-making based on factors such as who is entering an invoice. In spend management, AI can speed up time-consuming data entry by using optical character recognition (OCR) to ‘read’ documents, and natural language processing on financial document contents. AI assistance can also help employees understand spending policies, help managers ‘zoom in’ on anomalous spending patterns, and help controllers by suggesting classifications for transactions.
Worldwide, there are an estimated 14,000 regulatory changes each month across the approximately 19,000 tax jurisdictions. GenAI will be an invaluable tool for monitoring this massive number of tax content changes and be able to keep up with regulatory changes across internal compliance or audits. Over time, predictive models will even allow organisations to anticipate and get ahead of regulatory changes before they happen.
AI can be used to identify potential customers and geographies, driving top-line revenue while also helping bottom-line growth. For companies in the B2B space, AI fraud tools can cut detection times from days to minutes, meaning fraud detection can move from being reactive to happening in real-time. In the proof-of-concepts Sovos has piloted, we are already seeing accuracy up to 99%.
Effective Implementation for the Finance Function
For companies looking to introduce AI into their finance functions, there are best practices that can help make the evolution a success. From the outset, departments must have a clear understanding of the problem they are trying to solve, the value proposition and how feasible it is. Those making the decision need to consider the potential impact AI will have on stakeholders such as customers, employees, partners, and regulators.
As with most technology, the value of the results of using AI is only as good as the data put into it – and selecting the right tool for the job. AI is not a one-size-fits-all solution. There are many different types, such as machine learning, natural language processing, computer vision and deep learning. Finance and IT should work together to evaluate the availability and quality of their data, the scalability and security of their infrastructure and the compatibility and interoperability of their systems. Furthermore, not all large language models (LLMs) are created equal, and some will perform a given use case better than others. This is a great example of pilot often, fail fast, and iterate.
Ensure employees gain confidence in any AI solution by introducing it gradually, providing a solid foundation through training and education in how it can be used. There are three stages to AI adoption: algorithms making suggestions, algorithms offering shortcuts, and finally, algorithms doing the work, while expert employees monitor their progress. At each stage, business leaders need to assess how the AI solution is working to ensure it offers the greatest benefit to the organisation.
Stay Ahead of the Compliance Curve
Organisations that engage with innovative tax tech solutions that are already building AI into their technologies will gain a key competitive advantage. Generative AI frees finance leaders from tedious and time-consuming tasks and allows them to focus on strategic business functions, such as mergers and acquisitions, risk mitigation and more.
There are many ways generative AI can make life easier for tax and compliance professionals. Used with the proper preparation and with the right partners, generative AI empowers business leaders to streamline processes, reduce risk and enhance decision-making – allowing compliance to move from just a functional necessity to a true force for growth.
