Public Sector

Harnessing innovation: The technology transformation in tax administrations

October 15, 2024
Harnessing innovation: The technology transformation in tax administrations

Written by Paul McSavage
 

The global digitalization revolution

The rapid digitization of the global economy is drastically altering how we function at home and at work. Technology has reshaped business models, value chains, lifestyles, and daily work routines worldwide.

The rate of change is perhaps at its highest ever, with data at the heart of this transformation. While the overall volume of data generated, collected, duplicated, and used globally has been increasing significantly for years, only a small portion — 2% — of new data is presently retained . A thorough understanding of your data and the technologies available to interpret it are now essential to thrive in today’s data-centric world.

To keep pace with technological advancements, digitalizing global tax systems has now become imperative, with tax administrations’ modernization heavily focused on digital strategies to leverage the benefits of operating in the new data-centric world while minimizing its risks. Technology and digitalization can significantly boost tax administrations’ accuracy and efficiency in collecting, processing, and decision-making. Additionally, taxpayers can look forward to a more seamless and positive experience with tax compliance.
 

The modern tax administration challenge

The primary role of all tax administrations is to collect the correct amount of tax revenue, with governments relying on this income to fund critical public services. However, this is only part of the overall challenges and responsibilities they face. Public opinion has shifted following numerous global tax administrations successfully challenging the questionable historical “profit-shifting” activities of certain large, high-profile multinational groups and several high-net-worth individuals’ “tax optimization” strategies. While this is generally a positive development, it places more pressure on tax administrations to act fairly, efficiently, and effectively to maintain public trust and confidence in the tax system by proving that everyone pays their fair share of taxes. The Organization for Economic Cooperation and Development introduced the term “justified trust”   to describe tax administrations' work in building and maintaining community confidence that taxpayers are paying the right amount of taxes. Achieving this trust is more demanding and resource-intensive than simply confirming that certain risks do not arise.

Meanwhile, businesses continuously adapt for better economic returns as complex models, international operations, and global supply chains evolve amid increasing competition. Consequently, the complexity and scale of tax system administration have reached unprecedented levels, especially in a world demanding greater transparency and accountability.
 

Data is the key to success

When managing tax compliance, tax authorities try to make it as easy as possible for taxpayers to get things right while also identifying potential tax risks to address non-compliance. Data and information have become powerful tools for tax administrations, with significant improvements in both quality and quantity in recent years.

Harnessing innovation

While the increase in the quality and quantity of data should theoretically boost tax administrations’ risk assessment firepower, realizing the full benefits is not necessarily straightforward. Each tax administration is at a different stage along the technological development journey. Through collaboration with leading tax administrations, Moody’s has found that each has its own legacy systems, infrastructure, resources, and priorities, resulting in different end solutions. However, there are recurring themes among all tax administrations undergoing digital transformation, particularly in linking datapoints, utilizing data analytics, and overcoming automation hurdles. These challenges are precisely where cutting-edge technology, including artificial intelligence (AI), can help significantly.
 

AI in tax

AI is a catchall term that covers a number of technologies that enable computers and machines to simulate human learning, comprehension, problem-solving, decision-making, creativity, and autonomy. Machine learning, for example, involves teaching machines to discover relationships hidden in data to detect, predict, and make decisions based on certain criteria — which is particularly useful in the tax world. Deep learning uses neural networks to model real systems by mimicking how the human brain works. Most recently, generative artificial intelligence (GenAI) has arrived on the scene — technology that can create new content such as images, text, audio, and video based on the data it has been trained on using exceptionally large neural networks called large language models.

AI offers numerous advantages to the tax world, including:

  • Tax compliance — AI has the potential to automate processes to streamline data capture, computation, and output generation. GenAI can significantly reduce time-consuming aspects of tax compliance with further gains expected as the technology improves over time.
  • Tax controversy — AI can piece together disparate pieces of information and assess these against certain parameters to vastly improve tax risk identification and assessment. Machine learning can interrogate numerous datapoints and use algorithms to identify high-risk activities.
  • Transaction analysis — AI can improve the time-consuming process of analyzing vast amounts of transactional tax data and presenting findings and recommendations based on tax risk criteria.
  • Real-time compliance — AI can process data, provide analytics and recommendations, and act in near real time.
  • Process automation — AI can replace manual, repetitive, and time-consuming processes.  


At present, most tax administrations envision AI tools being used as a copilot, helping support tax professionals’ knowledge and judgment rather than replacing them altogether. However, technologies like AI are still in their infancy, so with eventual refinements in the underlying technology, trust in the reliability of their outcomes, and the right governance in place, full AI automation for certain aspects of tax administration is not out of the question. 
 

AI in action 

Although the tax industry is just beginning to incorporate AI, many global tax administrations are already realizing its benefits:

  • The United States Return Review Program — The US Internal Revenue Service (IRS) has been a long-term proponent of using data and analytics to help in the tax administration process. The IRS employs a combination of traditional and AI techniques, filters, and models to assess and identify potential noncompliant tax returns. The sheer volume of tax returns that can be reliably screened for potential follow-up action would not be possible without the use of AI, with advances in AI technologies likely to make the process even faster and more accurate.  In its first year of operation alone (in 2017) the IRS calculated that it prevented US4.4bn  of bad refunds.  
  • Australia's tax return prompting — The Australian Taxation Office is one of the most advanced tax authorities in the world, particularly in its use of data, analytics, and AI to help it achieve its goals. For example, it uses real-time data analytics and “nearest neighbor methods” to quickly assess amounts being claimed in electronic tax returns, with users receiving a friendly prompt to reassess any amount that doesn’t correlate to that expected. In 2020, approximately 340,000 taxpayers received such prompts, resulting in self-adjustments totaling approximately AU$37 million .
  • France’s improved detection — The French Tax Authority has recently been successful in using satellite imagery and AI to detect an additional 20,000 undeclared swimming pools, resulting in additional property tax revenue of €10 million .
  • Transfer pricing (TP) benchmarking — Tax authorities worldwide must consider the implications of AI solutions being used as part of the TP benchmarking process and results that support taxpayer outcomes. For example, we at Moody’s have developed an AI-enhanced version of our TP Catalyst solution that uses AI to significantly decrease the time taken to review and determine the most appropriate comparable companies used to price intra-group transactions.
     

Proceed with caution

As with any emerging technology, there are usually several legitimate concerns and possible obstacles to adoption that must be tackled to fully unlock their potential. In GenAI’s case, there are three key areas of concern for tax administrations:

  1. Hallucinations — Due to the nature of GenAI technology, it provides an answer regardless of accuracy, presenting it convincingly enough to seem believable.
  2. Lack of transparency — If there is inadequate training or lack of specific knowledge and processes, a GenAI-generated response may lack clarity regarding the analysis performed and sources used.
  3. Data protection — EU tax authorities must comply with the General Data Protection Regulation when using AI algorithms. This includes anonymizing personal data, obtaining consent, limiting use to specific purposes, and increasing transparency in data-processing.


None of these areas of concern are insurmountable on their own but they do need to be tackled as key concerns as global tax authorities begin to implement GenAI solutions.
 

The future?

Looking ahead, GenAI solutions will likely initially handle more routine, repetitive processes where the benefits and efficiency gains can be significant. As technology advances and lessons are learned from initial implementations — and with proper safeguards in place — the tax industry might be close to utilizing all forms of AI for real-time tax compliance, leading to a truly transformative tax system. Throughout this potential transformation, we must consider how it will affect humans’ future role in tax administration. As certain aspects of the tax system become automated, the skill set required of tax professionals is likely to shift, prioritizing problem-solving, storytelling, and judgment to complement the benefits AI will bring to achieve better outcomes for tax administrations.
 

Conclusion

Although we are in the early stages of the global digitalization revolution, it's evident that AI and digital technologies can improve tax administration by handling increasing amounts of data and automating tasks with improved decision-making. Global tax authorities must continue to adopt and invest in these technologies, recognizing their limits and making sure they complement essential human skills. 

Moody’s collaborates with over 80 of the world’s leading tax authorities, enhancing their systems with our extensive suite of tax-focused data and solutions. With the right data, analytics, and technology tools, tax administrations can more effectively manage their systems and achieve their goals in today's increasingly complex global landscape.

Learn more

Public sector solutions

Moody’s offers deep data intelligence, extensive publicly and commercially available information and insights, and predictive analytical tools that provide quicker and more efficient government decisions.

Learn more