AI is helping junior tax practitioners work faster and take on more complex tasks. But reducing the difficult parts of tax work may also affect how professional judgement develops. Junior practitioners...
AI is changing how junior tax practitioners research and learn. The challenge is ensuring they can check sources, question conclusions and understand how an answer has been reached. AI has made tax knowledge...
AI Is Already Embedded in Tax Work AI is already inside the tax team. It is in the first draft of the client email, the research summary, the document review, the knowledge search, the memo structure...
Loyalty starts long before someone resigns The resignation rarely starts on the day someone hands in their notice. In tax, it often starts much earlier: another late night before a deadline, another...
A quieter market, but not an easier one On paper, UK hiring should now feel easier. Vacancies have fallen, candidate availability has improved, and the intense post-pandemic recruitment market has cooled...
AI can help tax practitioners explore issues, structure analysis and identify sources. The strongest results come when it supports professional judgement rather than replacing it.
AI is already part of everyday tax work. Our February 2026 survey of 446 tax professionals from the UK and Ireland found that 84% use AI for tax research. It is also widely used for knowledge management (68%), document analysis (67%), drafting internal communications (65%) and drafting client documents (64%).
The benefits are clear. Among those using AI, 53% said it is helping them work faster and 32% said it is improving the quality of their work.
The more difficult question is how tax practitioners should use it.
AI can provide a quick, confident and well-structured response to a technical question. But treating that response as the finished answer creates risks, particularly for junior practitioners who may not yet have the experience to recognise missing qualifications, weak sources or incorrect assumptions.
Almost two-thirds of respondents, 63% said junior practitioners should instead be taught to use AI as a “thinking partner”.
That distinction can shape whether AI strengthens professional judgement or gradually replaces it.
AI works best when the practitioner has already thought carefully about the issue.
Before asking a tool for help, tax professionals should understand the client’s circumstances, identify the relevant questions and consider what information may be missing.
Adrian Hextall, Director at S&W, explains why that groundwork remains important:
“Knowledge will always be key to innovation and clear thinking. Knowledge also ensures an understanding when faced with business issues. AI might be the answer to a problem, but the problem must be understood first.”
Without that understanding, it is easy to ask a broad question and accept a broad response. The tool may produce something plausible, but it may not address the particular facts, risks or commercial priorities involved.
Using AI as a thinking partner means beginning with the practitioner’s own view. The tool can then help test that view, identify gaps and suggest alternative approaches.
Turn complex tax work into clearer, stronger outcomes
AI does not need to provide the first or final answer.
A practitioner could begin by setting out their own interpretation and asking the tool to challenge it. They might ask what assumptions have been made, which alternative interpretations should be considered or what additional information could change the conclusion.
This approach turns AI into a source of questions rather than a source of authority.
It can also help practitioners explore unfamiliar issues without handing over control of the analysis. AI may point towards relevant legislation, guidance or case law, but the practitioner still needs to assess those sources and decide how much weight they carry.
Jane MacKay, Tax Partner at Crowe, says this distinction has informed how her firm approaches AI:
“We’ve focused on guidance, training and positioning AI as a first-pass tool rather than an answer engine.”
A first-pass tool can help make the early stages of a task more manageable. It can structure a research plan, summarise competing positions or suggest where further investigation is needed.
The final judgement should still belong to the tax professional.
The most useful AI response is not necessarily the one that sounds most certain. It is the one that can be checked.
Practitioners should ask for the sources behind a conclusion, the assumptions used and any areas of uncertainty. They should also ask whether the answer depends on facts that have not yet been established.
This is particularly important because 74% of survey respondents said they are concerned about inaccurate or fabricated outputs, while 53% are worried about over-reliance.
More than seven in ten respondents, 71%, said they feel more confident using AI grounded in tax sources. Access to trusted material makes it easier to trace a conclusion and understand how it has been reached.
Even then, the sources need to be opened, read and assessed.
As Hextall explains:
“It may be able to do the research for you, but the results must be proofed before they are used.”
A citation is a useful starting point. It does not remove the practitioner’s responsibility to verify the answer.
A polished AI-generated document can make weak analysis appear more complete than it is.
This is particularly relevant for junior practitioners. The report found that 66% of respondents see verification and source-checking as their biggest skills gap, while 58% identified deep tax reasoning and argumentation.
Managers should therefore review how a conclusion was reached, not only the quality of the finished wording.
Junior practitioners could be asked to explain which sources they relied on, where AI contributed, which assumptions they tested and why they selected one interpretation over another.
This allows AI to support the learning process without hiding it.
Ian Bowden, Tax Partner at BDO, says the priority should be preserving understanding even where technology performs more of the practical work:
“The key is intentional design: we must teach the concepts even if the machine does the mechanics.”
AI can help practitioners work through a problem, but it cannot take professional responsibility for the conclusion.
The tax professional still needs to decide whether the answer is reliable, how it applies to the client and what advice should follow.
Using AI as a thinking partner means allowing it to contribute without allowing it to lead. It can suggest, question, compare and organise. The practitioner must still verify, interpret and decide.
That is likely to become one of the most important skills in tax: not simply knowing how to obtain an answer from AI, but knowing how to use the tool to reach a better answer of your own.