AI-enabled tax transformation

Notwithstanding the hype, many tax departments remain hesitant about embracing AI, caught between the excitement of technological innovation and resistance to change. Tax leaders, understandably wary of potential pitfalls (as inaccuracy and data security chief among them), are carefully weighing the risks against the promise of AI-powered efficiency.

Many companies have started simply with gaining familiarity with AI tools, from leveraging generative AI to boost productivity, to automating data-intensive tasks. According to Deloitte’s Tax Transformation Trends research, 21% of tax leaders are testing the water by prioritizing the automation of routine data tasks using AI (Fig. 2). Some have already started experimenting with more strategic tasks for AI such as tax planning enhancement (10%) and identifying tax risks and opportunities (8%).

With so much choice in AI tools on the market, and more being developed by the day, it’s important not to forget first principles and have clarity about the business or tax department objectives. Success factors could include streamlined compliance, enhanced insights, reduced costs, or improved accuracy. Clearly defined goals clarify the evaluation criteria for AI tool functionality.

 Typically, it is important to integrate AI tools into the operations where people spend their time. Also consider integrating data flows from Systems of Record/Enterprise Resource Planning systems and boundary systems into AI tools. For example, 46% of the survey respondents already enhanced their ERP systems with AI-driven tax analysis tools, and a further 44% plan to do so (Fig 3).

By prioritizing demonstrable business outcomes, tax departments can effectively leverage AI to achieve strategic goals.

With the right quality of data, AI can process information much faster than it can when dealing with poor data, providing quicker insights and allowing for more timely business decision-making. This necessitates a well-defined data strategy and approach, which is crucial for understanding what data should and could be used to advance AI-fueled business outcomes.

This strategy must also take into account data privacy and confidentiality, clearly defining what data is crucial for your tax process flows and identifying the systems that house that data. By meticulously preparing and cleaning tax data, a tax leader can unlock accurate insights, improve efficiency, and reduce risk.

By meticulously preparing and cleaning tax data, a tax leader can unlock accurate insights, improve efficiency, and reduce risk.

The greatest barrier to AI adoption in tax departments is trust, with 77% of tax leaders requiring 90% or higher accuracy before entrusting AI with their tax processes. This reflects the exacting and regulated nature of tax, its impact on the business, and a need for greater transparency and confidence in AI’s capabilities.

Beyond trust issues, other factors are preventing tax leaders from diving in headfirst (Fig 4). These include budget constraints (cited by 45% of Deloitte’s survey respondents), limited AI expertise within teams (36%), and a lack of a clear AI strategy (33%). Furthermore, concerns about data security and privacy (30%) and insufficient support from company leadership (28%) are significant hurdles.

While there aren't any quick fixes, these issues are solvable. A cautious approach is to start small, learn quickly, and build confidence along the way. Crucially, a programmatic and thoughtful deployment strategy, focused on solving core business problems, yields better engagement and ROI than simply introducing tools without clear guidance.

For most organizations outside of the AI industry itself, AI is new territory. A barrier to implementation (figure 5) is limited AI expertise in the team, and 45% of Tax Transformation Trends 2025 survey respondents identified AI-related skills as their greatest need in the next one to two years. In addition, 94% of respondents believed AI skills will be essential within the next four to five years (Fig. 5). In a bid to improve their tax teams' technological expertise, 53% of respondents said they are actively recruiting digitally native talent skilled in technology and AI. To address challenges, basic first steps should therefore include defining the outcomes you want to achieve with AI and planning to bring new skills into the team.

Given the wide distribution of tax work and stakeholders who will have a broader enterprise view on the specialist or supporting functions, talent strategy, and skillsets, fostering close collaboration with Finance and IT is essential for better use of AI across the enterprise.

For a critical evaluation of existing tax processes, tax leaders can choose to automate existing processes (a reasonably basic re-imagining), target specific pain points such as error-prone parts of processes (problem-driven), or a re-imagining from the ground up (outcome-driven).

Data is the fuel for AI. Respondents to Deloitte’s Tax Transformation Trends 2025 research cited integrating tax data across the organization (30%) as a top three issue, as well as limited tech/data management expertise (28%). For organizations with limited human resources, piloting AI-driven data management could be a practical first step.

Once processes are mapped and data foundations are addressed, tax leaders should evaluate the options for developing their AI roadmap. These could include leveraging existing Finance AI tools, build/buy/co-investing for dedicated solutions, running pilots with third parties, or outsourcing.

Low enterprise AI adoption and low tax authority AI implementation

This scenario assumes a world where tax authorities remain at a Tax Administration 2.0 model and the company’s tax team has not embraced AI beyond basic GenAI tools. Tax departments continue to rely largely on limited systems integration, using AI selectively for productivity or simple automation. Tax departments that see this scenario as likely in the medium term may choose to treat it as a transitional or legacy state.

High enterprise AI adoption and low tax authority AI implementation

In this scenario, the tax department has advanced significantly with AI adoption and routinely uses AI agents, but the tax authority remains at Tax Administration 2.0. Organizations will have embedded AI across finance and core business processes, while tax authorities lag behind. This scenario presents an opportunity for tax leaders to build a compelling business case for investment in tax-specific AI solutions and governance.

Low enterprise AI adoption and high tax authority AI implementation

In this scenario, tax authorities have moved to a real-time, data-driven model that expects upstream data quality and integration, while the company’s tax function has not adopted advanced AI and remains limited in automation. The imperative would shift to implementing proactive, AI-driven preventive controls upstream in business processes.

High enterprise AI adoption and high tax authority AI implementation

In this scenario, both tax authorities and companies operate in an AI-enabled, largely automated environment. Compliance shifts toward continuous or near real-time reporting, and the focus moves from detection to prevention. Human judgment remains instrumental. Tax professionals would transition from preparation to oversight, interpretation, policy interaction, and dispute resolution, embedded within enterprise decision-making rather than operating at its margins.

It is likely that many companies will find themselves in a combination of scenarios. Decisions that apply across all four include:

• Data integrity and common data models
• Clear ownership of AI-driven tax outcomes
• Explainability and audit trails
• Cross-functional integration with Finance, IT, and Operations

Some decisions, however, need to remain flexible.