How Tax Authorities and Taxpayers Can Use AI Tech in Compliance

“Tax compliance means adhering to government rules and regulations for paying and reporting taxes. It includes calculating and paying taxes accurately and promptly, maintaining records, and submitting necessary tax forms to the appropriate tax authorities.”

This definition by ChatGPT 4.0 covers many, but not all, aspects of tax compliance, which is more than just how tax laws are applied. It omits communication, training, risk management, and the economic benefits of compliance. These elements are interconnected and involve both tax authorities and taxpayers.

The goal is to achieve accurate tax compliance at the lowest possible cost while avoiding artificial and aggressive tax planning. Companies aim to optimize their process costs and minimize their tax burden, as long as it aligns with tax compliance principles.

Artificial intelligence techniques such as machine learning and process mining can automate various tax compliance tasks.

Solutions Based on AI Technology

Due to the increasing requirements for tax compliance transparency, for example new regulations such as the OECD’s BEPS 1.0 and 2.0 and the EU’s Anti-Tax Avoidance Directive and Directive on Administrative Cooperation, most tax administrations in the EU have started using solutions based on AI technology, and in three key ways:

The first group of AI algorithms performs a function of taxpayer assistance. Usually, such a system uses natural language processing to conduct a textual online conversation and answer taxpayer queries, substituting for human tax officers. With natural language processing, the chatbot recognizes keywords and uses decision trees linked to frequently asked questions to give feedback to the taxpayer’s query.
The second category of AI algorithms performs automated data collection functions. These are tools such as web crawlers or web-scraping systems that are used to collect taxpayer data from webpages, social media, e-commerce, or e-sharing platforms. Data obtained is matched afterwards with data already stored in databases of the tax administration.
The third category of AI algorithms is used to automate tax risk-detection functions. Social network analysis, transmission network analyzer, and anomaly detection are tools applied. The European Commission employed transmission network analyzer in 2019 as a tool to help combat value-added tax fraud. Anomaly detection uses AI algorithms to perform checking of unusual discrepancies between reported (or unreported) data and the available information.

Taxpayers’ Rights

The risk management algorithms are used to predict tax risks or non-compliance with obligations. This categorization is an important issue regarding the question of how EU member states should structure their legislation regulating AI algorithm applications to achieve the necessary balance of rights and obligations between taxpayers and tax authorities.

For instance, the use of chatbots based on AI algorithms by tax authorities probably wouldn’t pose the threat of limitation of taxpayers’ rights, as practical exploitation of the function of assisting taxpayers wouldn’t infringe their rights or obligations. Therefore, information gathering, risk management, and supporting systems are the focus of tax administration compliance.

Analyzing Business Processes

AI can be categorized into two orientations—data-driven and rules-based.

Data-driven AI uses prediction models, natural language processing, visual law, and network analytics to compute tax laws. Rules-based AI refers to expert systems, self-executing law, and computable codes. These tools focus on automating tax processes based on taxpayers’ objectives in specific use cases.

To analyze business processes, such as those triggered by a business transaction, it is useful to employ business process modeling notation, or BPMN, to create visual models. Process mining techniques can also be used to uncover the actual processes, helping to ensure tax compliance.

While traditional workflow systems with activity log or audit trail functionality are well-suited for BPMN, other systems such as enterprise resource planning, supply chain management, customer relationship management, document management systems, custom applications, and related systems can be used to identify tax-related process activity.

Consequently, the following process mining approaches are of particular interest:

Value-added or sales tax: The process of an order via goods receipt, invoice receipt, tax invoice verification, and payment of sales tax after a tax return has been prepared, will be accessible to a process mining analysis.
Customs: The movement of goods during import and export for purchases, sales, and intra-company movement of goods between affiliated companies provide interesting starting points for analysis. In particular, the combination of the sales tax, VAT and customs processes leads to new insights into the quality of the processes.

Key Lessons for Business

Recognize the role of AI in tax compliance. Tax administrations use AI technologies, such as natural language processing, data collection tools, and risk-detection algorithms, to improve tax compliance processes. Business leaders should be aware of AI’s role in tax compliance and its pros and cons.

Consider AI as a tool for taxpayers’ assistance. Taxpayers can receive assistance with tax compliance through the use of AI-powered chatbots. Additionally, businesses can automate data collection and communication with tax authorities by using AI tools such as web crawlers.

Use AI for tax risk-detection. AI algorithms such as social network analysis and anomaly detection can help in identifying potential tax risks and non-compliance. Business leaders should explore the application of these risk management algorithms to predict and prevent tax-related issues within their organizations.

Visualize and analyze operational processes. Using BPMN and process mining can improve tax compliance by identifying areas for improvement and ensuring accurate reporting.

Explore automation in tax processes. AI can help businesses automate tax processes such as transfer pricing and payroll tax, boosting efficiency and accuracy.

This article does not necessarily reflect the opinion of Bloomberg Industry Group, Inc., the publisher of Bloomberg Law and Bloomberg Tax, or its owners.

Author Information

Robert Risse is a long-standing faculty member of the WU Executive Academy, the business school of the Vienna University of Economics and Business (WU), and director of WU’s tax law technology center. He was corporate vice president of tax and trade at Henkel AG for over 20 years.