Keep the knowledge, control the costs: how tax departments can use AI to manage rising obligations

This is a publication that is part of RSM's Voice of SCM. On a monthly basis, RSM issues the Voice of Supply Chain Management (SCM). Our SCM consultants constantly follow global SCM developments in an ever-changing society and translate their impact into practical considerations for internationally active companies. 

Tax directors of internationally active companies face new reporting and disclosure duties year after year, often with little change in team size. Neither automation nor reskilling alone will close that gap. Hiring, outsourcing and new tools can each help, but each brings costs and trade-offs. Outsourcing can gradually move the department's knowledge outside the company. We believe tax departments need a layered approach: build and maintain their own knowledge base, then use AI and outside expertise to put that knowledge to work.

Six months ago, we published our white paper AI & Tax: we have got a story to tell on the technology side of this question. Since then, the obligations have kept coming, the promised simplification has stayed on paper and our conversations with tax directors have sharpened our view. The white paper described the machine, and the question now is who runs it and how a tax department grows its capacity when the team itself does not grow.

This article was written by Mario van den Broek (MvdBroek@rsmnl.nl) and Mourad Seghir (MSeghir@rsmnl.nl). Mario and Mourad are consultants with RSM Netherlands with a focus on international tax, transfer pricing and tax technology.

Growing obligations and limited capacity

Why tax reporting obligations keep increasing

Over the past decade, governments have worked together to make tax more transparent, and the tax department of every company active in international trade has carried the result. The OECD Base Erosion and Profit Shifting (BEPS) project brought country-by-country reporting (CbCR) and the master file and local file for transfer pricing. The European Union added a series of Directives on Administrative Cooperation (DAC), including DAC6 on reportable cross-border arrangements. Pillar Two introduced a global minimum tax with its own GloBE Information Return (GIR) and national top-up tax returns. E-invoicing, real-time VAT reporting and Standard Audit File for Tax (SAF-T) requirements are spreading country by country. These are examples, and the list is longer.

These rules share a practical consequence: companies must collect, structure and report data that tax authorities used to gather themselves. They must do so for each entity and country, within fixed deadlines. Authorities then exchange and compare that data across borders. Part of the work of tax administration now takes place inside the tax department, without a corresponding increase in people or budget.

Governments have good reasons for this, and none of them is going away. Tax authorities want to close the gap between the tax that is due and the tax that is collected, and they have learned that data from companies is faster and cheaper than an audit. The OECD has described this direction as Tax Administration 3.0, in which tax processes move closer to the systems in which companies record their transactions. For the tax department, this means the authority increasingly sees the same data as the company, and sometimes at the same moment. Errors that used to surface years later in an audit now surface within a filing cycle.

The past six months have confirmed that direction. Most groups in scope of Pillar Two filed their first GIR and their first Dutch top-up tax return in the summer of 2026, and for many teams that meant weeks of collecting data that no system held in the right form. On 24 June 2026 the European Commission presented a Tax Omnibus and a recast of the DAC rules, with claimed savings for business running into billions of euros. The Council, however, only aims for agreement on the Omnibus by the end of 2027, and some measures would not take effect until somewhere between 2032 and 2037.

For now, simplification exists on paper, and when it does arrive, it will be one more change to implement next to the rules that stay in place.

Increasing market pressure

Why tax teams are struggling to keep up

AI tools have also developed over the same six months. They have become better at carrying out a sequence of steps, and more tax teams have started using them in their daily work. As the tools improve, departments need people who understand the output, check it and take responsibility for it.

The teams carrying this load have not yet grown to match it. Our white paper was written for central tax teams of five to fifteen people covering corporate income tax, transfer pricing, VAT and employer obligations for groups active in five to forty countries. The Thomson Reuters research we cited found that more than half of corporate tax departments already felt under-resourced, even before the first Pillar Two returns were due.

What tax directors need from their teams

Earlier this year we sat down with the tax director and team of the Dutch European headquarters of an internationally active company and asked what mattered most to them. They wanted to comply in every country where the group operates without paying more tax than the law requires. They also wanted to avoid positions they could not defend: any savings would be outweighed by the damage to their reputation and relationship with tax authorities. They needed to keep up with obligations as the business entered new countries and sales channels. With some reluctance, they added that they did not want to keep asking for more people or spend a fortune on external tax consultants.

It helps to look at what each goal means in daily work. Being compliant in every country means that a team in the Netherlands answers for filings in countries where it has no staff, often through local service providers whose work it can only partly check. Not paying too much tax means knowing which reliefs, treaty positions and incentives apply, and applying them every year on the department's own initiative. Not overstepping means that every position must be one the team can explain in its own words. Keeping up with growth means that a new market or a new online channel triggers registrations, filings and transfer pricing questions within months. The last two goals are about the standing of the tax department inside the company. A tax director who can show that the team does more with what it has earns the trust to ask for investment when it is needed.

These are reasonable goals, but meeting them together is difficult. We hear similar concerns from almost every tax department we speak to. The conversation reinforced our view that each option needs to be judged against all these goals: relieving one pressure can increase another. 

Keeping tax knowledge in house

The hidden cost of outsourcing tax knowledge

Outsourcing adds expertise without adding headcount, which makes it an attractive option. Its less visible cost is the knowledge that can accumulate outside the company. A memo may sit in an inbox while the reasoning behind it stays with the adviser. The following year, the department may pay to ask the same question again. When the adviser's team changes, part of the group's tax history can be lost.

Over time, the department can become a coordinator of advisers, less able to challenge their work or explain its positions to a tax inspector. That makes defensible tax positions harder to maintain. The team needs to understand the reasoning behind a position and retain a record of it.

How AI can help retain and reuse tax knowledge

AI can make a department's stored knowledge easier to find and reuse, provided the underlying records are current and access is controlled. Our white paper described a shared, maintained collection of tax data, memos, procedures, rulings and guidance. AI can help the team search those records in plain language and check answers against source documents. The next question can then start from the group's own reasoning rather than from a blank page or a new engagement letter.

The records need not be complicated. For each important position, keep the relevant facts, the rule applied, the reasoning, the decision and who took it, and a date for review. Our white paper recommended a tax library with version control, in which final documents are read-only and version history stays on. AI may make those records easier to search: a team member could ask why the group charges a certain royalty rate, or how a permanent establishment question was settled three years ago, and find an answer linked to the group's own documents. This also makes the cost of lost knowledge more visible. Every engagement either strengthens the department's own records or leaves it dependent on someone else to explain the group's tax position. Tax directors can test that dependency by listing the questions the business asked over the last year and checking how many the team could answer from its own records without calling an adviser. That share is a useful baseline for improvement.

Ask what the department will retain from each advisory engagement. Advisers should deliver their conclusions, reasoning and decisions in a form the team can store and use, with a short session to explain them. A report alone may leave the same questions to be answered again next year; a documented handover makes the work useful beyond the immediate engagement.

Building capacity

Five layers of tax department capacity

Our white paper examined the tax department from the bottom up. We take the same approach to capacity, with five layers: understanding the workload, building internal tax knowledge, training people, using AI and bringing in outside help where it adds most. Start with an obligation map covering the next three years. For each obligation, identify the data needed and the people responsible for preparation and review. This gives the resource discussion a factual basis.  

Many teams underinvest in their records and in the people who maintain them. Both are essential to reliable use of AI: outdated material and untrained reviewers can leave confident but unreliable answers unchallenged.

External advisers provide the fifth layer, contributing specialist judgement, support at busy times and a view of upcoming developments. Each engagement should leave the department better able to explain its own tax positions. 

What skills will future tax teams need?

When the budget allows for a new hire, start by identifying the skills the department needs. Some preparation and reconciliation work can be automated. People still need to specify rules precisely enough for software to apply them, review machine-produced output, maintain tax records and explain positions to the business, auditors and tax authorities. Hiring decisions should reflect this work, with part of the budget also reserved for learning and tools. These skills are not always available in one candidate. In our experience, graduates arrive with a solid grounding in tax law but little practice in working with data, AI or the commercial logic of the business. Departments will need to develop those skills themselves.  

We see growing value in tax professionals who understand how the business makes money and can use AI to connect that understanding with their tax expertise. In a company selling through multiple channels and countries, they can explain what a new online sales channel, marketplace or distribution hub means for VAT, customs and transfer pricing while the commercial decision is still being shaped. This is where supply chain management and tax meet.  

Over the next few years, we expect the balance of the tax team to shift further in this direction. As AI takes on more preparation work, fewer people will be needed to produce numbers and more to interpret them. Their work will include challenging business proposals early and explaining the tax consequences of changes to supply chains or sales models before they happen. Professionals with both tax and commercial skills can bring the department's knowledge into those decisions, making its advisory contribution more visible alongside its compliance work. Few candidates arrive with all these skills, so tax directors will need to develop them within their teams. Rotations into finance, supply chain or commercial teams can help, as can joint projects with colleagues designing a new route to market. AI tools can also help tax specialists explore data and business models without waiting for an analyst. Better business understanding helps them ask more useful questions of both the department's records and external advisers.

AI may also change a traditional training ground. Junior tax professionals have often built judgement by preparing returns, reconciliations and documentation; if tools take on more of that work, teams will need to create opportunities to practise reviewing it. Otherwise, they risk having tools that few people can challenge confidently. One answer is to build practice deliberately, much as airlines use simulators to train pilots for uncommon situations. Our white paper was itself built on simulations, in which we positioned ourselves as the client and ran complete tax cycles to see where time and quality were lost, and the same method can support training. For a tax team, useful cases include a tax audit that questions a transfer pricing position, entry into a new country, a restructuring, an unexpected Pillar Two top-up or a convincing AI output that turns out to be wrong.

The team can take a filing or process it knows well, write down the facts of a realistic case and a reviewed outcome, and ask an approved AI tool to create variations, such as a changed rule, a missing document or an unusual transaction. At least one case should contain a convincing but wrong answer. Working through those cases together, with a discussion of where people went wrong, gives the team a way to practise judgement; the exercises should be checked by a tax specialist and use appropriate data safeguards. Once a real case is closed, its facts, reasoning and outcome can be documented and, where appropriate, adapted into a future exercise. New team members can then learn from the group's own decisions, with AI helping them locate the underlying records. The same documentation can support both daily work and training.

Tax professionals will also want to understand their career prospects as preparation work declines. We see senior opportunities in reviewing and signing off output, taking responsibility for the department's knowledge in a tax area and advising the business across tax and commercial questions. These roles offer scope for more judgement and responsibility.

Tax directors need to develop their own AI skills as well. First-hand experience helps them judge the team's output, set workable rules and lead changes in how the department works. In our practice, we set aside time each week to work on AI applications drawn from daily client work, one project at a time. These sessions have shown us how confidence develops through use. Tax directors can do the same by using AI to prepare a business meeting, test a position or summarise a new rule, then sharing what worked and what did not.

Putting AI to work

How to start using AI in a tax department

The next step is to choose where AI could make a useful difference in the team's daily work. Use the obligation map for the next three years to identify one recurring process that takes too much time or depends on knowledge held by only a few people. Start in a tax area the team knows well, so it can judge the results. Agree what AI will help with, which records it may use and who will review its output.

Bring the relevant facts, memos and previous decisions together in the department's knowledge base, then test the chosen process using an approved AI tool. For example, the team could ask it to retrieve the reasoning behind an existing tax position and prepare an initial response to a business question, with references to the underlying documents. A tax professional should check the sources, assumptions and conclusion before the response is used. Working through a familiar case makes it easier to see where AI saves time and where its output needs correction.

External advisers can help the team prepare and assess that trial. Their contribution might include checking the source material, challenging an AI-generated answer or helping colleagues recognise when specialist judgement is needed. Agree in advance what the adviser will deliver, how the reasoning will be recorded and how the team will learn to repeat the process. A fixed fee for a defined scope can give cost certainty, with room for questions as the team learns. Tax directors should also ask how the adviser uses AI and how any efficiencies are reflected in the proposed work and price.

The safeguards discussed earlier need to be part of the trial from the outset. Assign someone to keep the records current, decide which tools may access them and establish where data is stored and who can see it. Make responsibility for reviewing and signing off the output explicit. Protect time for learning as well as funding for tools: the team needs practice to recognise a plausible answer that is wrong. One of the simulations described above can provide a useful test before the process is used more widely.

After a few cycles, compare the time spent on preparation and review with the previous approach, including the effort needed to correct mistakes. Check whether more people can explain the underlying tax position and whether the team can answer more business questions from its own records. These measures give the tax director a basis for deciding whether to improve the process or extend it to another area. Where AI releases capacity, use some of that time to examine upcoming obligations and business decisions with advisers, while there is still time to act.

This brings the five layers together in a manageable first step: a known obligation, reliable records, people who can review the work, an AI tool and specialist support where needed. The department can build from experience, retaining what it learns after each cycle. Over time, that should help it handle more work while keeping the knowledge needed to explain and defend its tax positions.

RSM is a thought leader in the field of supply chain management including ESG. We provide frequent insights through training and thought leadership, based on detailed knowledge of industry developments and practical application gained from working with our clients. RSM helps companies connect individual geopolitical and regulatory developments to the broader changes affecting their supply chains. By doing so, we support businesses in distinguishing short-term disruption from structural change and translating that assessment into coordinated decisions on sourcing, origin, landed cost, logistics and supply-chain design. If you would like to discuss what these connected developments mean for your business, please contact one of our consultants.