Why AI is poised to make managing tax risk and controversy easier

The Tax Risk and Controversy survey provides evidence that tax functions are moving steadily along the GenAI adoption curve. The fact that 39% of respondents say they’ve built at least one tax risk or controversy-focused pilot or tool and 30% more have begun integrating GenAI into other key processes marks dramatic progress in line with the 2025 Tax and Finance Operations survey, which also finds functions advancing steadily across their GenAI journey.

Respondents generally say they’re using or anticipate using GenAI for three primary purposes:

• Analyzing and summarizing large volumes of external tax information (especially new and proposed legislation)
• Analyzing and summarizing large volumes of internally generated tax information (such as opinions, memos, meeting minutes)
• Enhancing the consistency of data shared with tax authorities

Significant proportions are also using it for other tasks such as better identification and management of nascent tax risks, generation of supporting evidence and documents in real time, and automating or partially automating responses to tax authority enquiries. Separately, a little more than a quarter of respondents say they’re using tax technology (but not necessarily GenAI) to perform predictive analytics.

Daren Campbell, EY Americas Tax Innovation Leader, says pilot projects should be considered in early stages of adoption.

“Most pilots and early use cases I’m seeing are related to classification, document summarization and key element extraction,” he says. For tax controversy specifically, “We have use cases around notice tracking and using AI to extract key elements from notices.”

How tax authorities are integrating GenAI

Tax authorities are also moving along the AI adoption curve. An Organisation for Economic Co-operation and Development (OECD) report on tax administration shows 29 of its 38 members are using some form of AI as part of their tax administration as of 2024; 79% of them are using it specifically to detect tax evasion and fraud, while others are using it for reasons ranging from virtual assistants for agency executives to improving customer service for taxpayers.¹ Use cases range from analyzing maps to detect unreported pools in Greece that should be paying property tax to sophisticated chatbots in Singapore that handle a range of corporate and income tax questions. 

As tax administrations become more comfortable with handling large amounts of data and advanced technology, they are adopting these applications at an accelerated rate. Between 2018 and 2022, the number of administrations using tools with embedded AI and machine learning (ML) technologies increased by 34%, and 8% more were planning to follow suit.²

Now, they are beginning to investigate and develop solutions that leverage GenAI to automate tasks such as tax return review, document auditing and providing taxpayer guidance. Dirección de Impuestos y Aduanas Nacionales (DIAN), the tax authority of Colombia, for example, is using GenAI to create personalized messages for taxpayers to increase compliance and reduce evasion. France is using it to combat property tax fraud. Japan uses it to analyze data from a variety of sources to determine which taxpayers are likely to not file returns, information it uses to audit more efficiently. 

Improved satisfaction

The survey shows business tax functions already integrating the technology are liking the results. Some 91% of those using AI of any kind say they’re somewhat or very satisfied with their controversy management, which is nine points higher than respondents overall.  

Moreover, those already integrating GenAI in tax controversy were far more likely to say they are “very” satisfied (46%) than those who have not (31%). And, among those already integrating GenAI into other key processes, tools or platforms, 92% agree or strongly agree that “companies that incorporate GenAI to create a single, integrated, end-to-end tax risk and controversy platform will have the greatest success in managing tax controversy in the future.”

“Everyone sees the potential, and the early movers are already starting to realize dividends,” says Coronado. “This will create a positive reinforcement loop that only encourages more innovation, which will help equip tax functions to operate confidently in this future state of tax controversy.”

To be sure, respondents face many challenges to further integration. Securing talent that knows how to work with the technology is the biggest barrier, cited by 44% of respondents. Securing budget and having insufficient insight or ideas around GenAI use cases are also obstacles.

“Integrating GenAI and agentic AI will be an ongoing process,” Campbell says. “The good news is it can deliver results almost immediately.”

GenAI won’t make tax controversy go away, of course. On the contrary, there are more sources of it than ever, primarily driven by rapid and relentless legislative and regulatory change. But GenAI is on track to emerge as a useful tool for managing risks and close preparation gaps even as tax functions develop technology to manage their existing controversies.

As in recent years, a combination of international cooperation on tax policy and tax administration will be responsible for a growing number of tax disputes in the coming three years, survey respondents say. Disputes may also take longer to resolve.

Concerns about Pillar Two and digital services taxes grow

On the policy side, future controversy will be driven largely by more jurisdictions enacting tax reform legislation recommended by the OECD because of its project on base erosion and profit shifting (BEPS). While many of these changes are being implemented on a country-by-country basis, only about half of respondents describe themselves as “highly prepared” to handle flash points related to many of them. 

Specifically: 

• 92% say growing implementation by individual jurisdictions of a global minimum tax as called for by BEPS Pillar Two will “somewhat or significantly” increase the number of disputes. 
• 91% say failure of countries to reach agreement on whether to reallocate a portion of the profits of the largest multinational enterprises (MNEs) to the jurisdictions where their customers are located, regardless of whether the MNEs have a physical presence there, will “somewhat or significantly” increase the number of disputes. 
• 90% say recommendations concerning transfer pricing will “somewhat or significantly” increase the number of tax disputes. Only 49% say they’re “highly prepared” to manage this coming workload.

Concerns about the global minimum taxes are omnipresent in tax functions, which are working to file the first Global Anti-Base Erosion (GLoBE) information returns in 2026, which calculate, among other things, whether top-up taxes are owed in certain jurisdictions. 

The fact that digital taxes were identified, for the first time in the survey’s history, as the most significant source of tax risk over the coming same three years suggests some tax executives are skeptical about whether agreement can be reached on Amount A, the new taxing rights in BEPS 1.0. Amount A reallocates a portion of profits of the world’s largest and most profitable MNEs to the market jurisdictions where their customers or users are located, regardless of physical presence. The result of no agreement could be new tax disputes over their digital activities. That could be bad news for the technology, media and telecommunications sector, which alongside private equity, reports having the most ongoing tax controversy of any sector; 12% of respondents from those sectors say they have more than 100 active audits.

The predictions for controversy related to Amount B, which aims to simplify routine transfer pricing distributions, are less surprising because managing transfer pricing controversy has long been a top concern for respondents to the survey (and transfer pricing was again a top five source of risk). However, the fact that only 50% say they’re highly prepared suggests an opportunity exists to implement more technology including AI to help. For example, 49% say they expect to tailor their transfer pricing documentation to the specific needs of individual tax authorities, a task that could be simplified with an AI agent.

The potential for GenAI to offer solutions for reducing risks associated with BEPS is being closely studied, particularly because it can be used to help organizations access and manage data. Beyond tracking developments with Pillar Two, it might prove helpful when specifically tailoring transfer pricing evidence and documentation collection processes to better meet the specific demands of one or more national tax authorities. It might also help businesses expand and improve their collection of contemporaneous transfer pricing evidence and documentation.

More tax transparency means more controversy

Current tax controversy is largely being driven by more tax audits, more demands for information, more transfer pricing adjustments and better coordination between governments. Some 99% of respondents say they have at least one tax dispute ongoing, and 69% say they have between 11 and 99.

Two developments in tax administration fueled by growing tax transparency may increase those workloads. Specifically:

• 92% say the exchange of information between tax authorities of different jurisdictions will “somewhat” or “significantly” increase the number of tax disputes.

• 92% believe some jurisdictions deciding to make business disclosures known as country-by-country reports available to the general public will “somewhat” or “significantly” increase the number of tax disputes.

The exchange of tax information between tax authorities has grown in tandem with broader cooperation between authorities, which may explain why one-third of survey respondents say they expect increasing coordination between tax authority departments in the coming three years. Much of the work is aimed at reducing double taxation or trying to give taxpayers more certainty in areas including Pillar Two compliance and to reduce filing burdens, but there is also more cooperation in enforcement matters. The Common Reporting Standard, which facilitates the automatic exchange of financial account information between tax authorities globally, has grown exponentially and now counts the participation of more than 120 jurisdictions. Separately, the OECD’s Forum on Tax Administration (FTA) Joint International Taskforce on Shared Intelligence and Collaboration (JITSIC), originally created in 2004 to combat cross-border tax avoidance, now has more than 30 active members. 

The public disclosure of CbCR is a newer phenomenon. The world’s largest multinationals have been required to provide country-by-country tax and financial information to a growing number of countries since 2016 as part of the original BEPS project. The report gives tax authorities more information about the business’s geographic footprints and how its tax and economic activities are related. Because much of the data is commercially sensitive, tax authorities were obligated to keep the data confidential. 

In the last year, the European Union (EU) and Australia have enacted legislation requiring companies to make those reports public. Under the EU directive, MNEs must disclose publicly the income taxes paid and other tax-related information such as a breakdown of profits, revenues and employees per country for all 27 EU Member States and all jurisdictions listed as non-cooperative jurisdictions for tax purposes. 

Multinationals now fear the reports may be misinterpreted, creating both tax and reputational risk concerns. This is one area where GenAI may prove useful – more than ever, multinationals are focused on reporting accurate and consistent data and that it aligns with other things they say about themselves publicly. The ability to use GenAI to quickly analyze their digital footprints and compare public and financial statements will help that effort. It might also help businesses with documentation related to potential tax refund claims.

E-invoicing and tax incentives 

Tax risks associated with tax incentives are likely an aftereffect from policies enacted as economic stimulus during the COVID-19 pandemic. The enforcement intensity for indirect taxes, especially value-added tax (VAT), is likely to increase as electronic invoicing scales.

Propelled most recently by the European Commission’s VAT in the Digital Age (ViDA) proposal, e-invoicing may provide the best glimpse of the future state of tax administration – including tax controversy – because taxpayers are increasingly interacting in real time with tax authorities. The benefits to governments include detect patterns, forecast future trends (i.e. criminal behavior) and reducing fraud. The benefits for taxpayers include streamlining compliance by improving accuracy, business efficiency and low lead times. Both parties stand to benefit from efficiencies in the process.

The tax authority’s ultimate goal is to develop an enhanced electronic database (progressively automate risk assessments, audit selection and tax collection by leveraging electronic documents) for everything in the supply chain, starting from goods production that is capable of fully automating examinations and audits.³

But the logistics of dealing with a labyrinth of rules and regulations and constantly shifting implementation timelines are becoming more complex as more jurisdictions adopt the standard (effective measures include exemptions from selected formalities during onboarding, linking consumer e‑receipts to income‑tax deductions, and pre‑filled VAT returns supported by free web/mobile issuance tools). Integrating new e-invoicing requirements into existing IT platforms is challenging for large multinational enterprises. Taxpayers are also having to apply reverse engineering on the interpretation made by tax authorities when proving upfront tax calculations and returns that demand payment in conflict with taxpayer positions.

Here, too, GenAI’s data retrieval and categorization capabilities may offer some solutions and enable personalization of services. 

The new technology-enabled tax controversy environment will require new approaches and new strategies. With more tax controversy looming, the survey finds businesses overall are still unprepared for the future, and many are still in a reactive posture. Only 31% say they are “very satisfied” with the way in which their company currently manages tax controversy. The most common reasons for dissatisfaction include a failure of local tax and finance personnel to effectively manage or report up new local tax controversies, “the attitudes of tax authorities toward large business taxpayers” and the volume and complexity of new or evolving tax laws and regulations.

Moreover, just 9% of respondents say they have “complete visibility” over their ongoing tax audits and disputes, which is sharply lower than past surveys, when it has been about a quarter. (To be sure, 49% say they have “substantial” visibility.) Notably, those with full visibility of disputes are more likely already to manage controversy with a global audit and dispute tracking technology or platform (34%) compared to those without full visibility (25%). They’re also more likely to be very satisfied with how their company currently manages tax controversy (67%) compared to those without full visibility (28%).

Most respondents say they want to take a more proactive approach. Ninety-one percent of respondents are “somewhat or significantly” likely to increase their focus on improving tax governance globally in the next three years.

Better tax governance means better tax controversy management

Strong tax governance will be crucial to face the future of tax controversy because tax authorities regard it as a key indicator of a group’s tax affairs.

At its core, robust tax governance focuses on creating a culture of accountability, transparency, prevention and ethical behavior. While technology, processes, and documentation all play pivotal roles, they are but tools in the hands of the people responsible for both overall tax approach and daily tax operations.

Operating in such a dynamic tax environment, where tax considerations are continually changing and shifting, may mean that many frameworks are not keeping up with the pace of change. They require repeated updates, potentially resulting in pockets of risk, which render the framework unsustainable. The use of GenAI may help taxpayers to keep up with the rapid pace of change by not only identifying changes to the tax environment, but also by facilitating the process of updating tax governance frameworks. The most effective modern governance models combine specific tasks with a set of overarching principles through which such tasks should be viewed and executed and are adaptable, allowing the governance framework to inform decision-making when new circumstances are presented.

Closing the preparation gap

The survey finds that nearly nine in 10 of the largest businesses – those with revenues of more than US$100 billion – say they have implemented one key tax governance activity to assess tax risk such as putting in place a tax risk “committee.” But almost none have implemented all of them. 

Smaller businesses were also more likely to have implemented at least one of these activities. For example, more than seven in 10 businesses with revenues between US$10 billion and US$99.9 billion have implemented at least one activity. And none have implemented all of them. This suggests progress is slow in implementing strong tax governance.

Respondents who use a systematic approach with one or more of these activities are more satisfied with their overall management of controversy. And about two-thirds of those who have not implemented these activities acknowledge doing so would “significantly improve” their outcomes. 

Alternative dispute resolution

Beyond strengthening tax governance internally, it will become more important than ever for many businesses to build trust and confidence with tax authorities. This includes making use of pre-filing programs that offer a greater degree of certainty and post-filing assistance intended to help manage double taxation when it does occur. Here, the survey finds more opportunities for GenAI to potentially help, as respondents express a lot of frustration with the way these current programs work, and few are using them successfully.

Advance pricing agreements (APAs), for instance, in which transfer pricing is agreed to by taxpayers and governments before returns are filed, are one such tool, are described as hard to secure because they take too long, cost too much or are unavailable in desired markets. As a result, just 25% say they have a defined APA strategy. It is possible GenAI, with its ability to analyze large amounts of data and make recommendations, could eventually help perform arm’s length analysis to determine proper pricing. 

Respondents also expressed some frustration with the Mutual Assistance Procedure (MAP) administered by the OECD. MAP offers post-filing assistance when double taxation results from cross-border disagreements. Respondents who use MAP wherever and whenever possible are slightly more satisfied with their overall management of controversy (85%) compared to those who do not (80%). However, 48% say they were denied MAP by at least one jurisdiction, and 46% say MAP is too complex or long in duration to consider using it.

Tax controversy is about to be forever changed. The days of resolving disputes over positions taken years or even decades before are giving way to real-time compliance. This is already happening around indirect taxes, and more tax transparency is nudging stakeholders in the same direction for income taxes. 

Technology, especially GenAI and agentic AI, is emerging as a critical tool for both taxpayers and tax authorities. But like all tools, they are only as useful as their user. To benefit, tax functions, especially those that focus on tax risk and controversy management, should be actively transforming their approach, so they’re in the best position to succeed in the future state. Here are four things businesses should be doing now:

Embrace and integrate GenAI and tax technology 

Tax functions have been at the forefront of business use of GenAI since its capabilities became widely accessible in 2022 because they are the clearinghouse for all of a group’s data. GenAI tools already are changing how tax functions work, and more innovations will come, especially as data scientists focus on use cases to solve perpetual controversy management challenges. To the degree GenAI can be used to establish common facts and data, it will bring more accuracy and efficiency to dispute resolution. It will also help facilitate the real-time compliance that is increasingly being demanded. Businesses cannot afford to be left behind.

Get better visibility over tax disputes

Total visibility overall tax disputes has long been an objective, but the goal has taken on more urgency with widening implementation of Pillar Two global minimum tax rules at the individual country level and concerns that unilaterally imposed digital services taxes may become new risks that need to be managed. The survey results are indisputable: Tax functions with greater visibility over all of their controversies have more satisfactory outcomes.

Strengthen tax governance

Having strong, adaptable tax governance both signals credibility to tax authorities who rank taxpayers according to risk and positions businesses to respond confidently when disputes do happen. This is the “humans in the loop” component that complements and helps manage the growing influence of GenAI in the tax controversy environment. When it comes to adopting the seven key elements of good tax governance, businesses should deploy all seven key activities; too many have deployed only one.

Seek certainty wherever possible using alternative dispute resolution programs

The frustrations with APAs, MAP and other alternative dispute resolution programs have been well documented throughout the history of this and the EY Transfer Pricing survey. However, the combination of stronger tax governance and deployment of GenAI might help make qualifying for those programs more efficient or possibly easier. It will be important in the future state of tax controversy to pursue any certainty that can be found. 

Summary

The 2025 Tax Risk and Controversy survey finds tax functions are investing in AI tools with the expectation they will improve management of risk and controversy. This is especially true as sources of tax controversy diversify, including ongoing concerns about Pillar Two global minimum taxes, transfer pricing and the emergence of digital services taxes as a top concern. Strengthening tax governance is also a crucial step for future success.