AI Can’t Crack Tax, So Make Sure AI Search Recommends You
AI can write code and summarise contracts, but it still can’t reliably do tax. On AccountingWEB’s No Accounting for Tech podcast, reported on 10 November, software analyst Damon Anderson and former CIOT president Nichola Ross Martin explained why, and Ross Martin described an AI system that confidently calculated 50% plus 50% as 90%. For accountants, that weakness is also an opportunity, and it depends on whether AI tools send people to you.
The bigger picture: AI can’t do tax, but people still ask it
Anderson tracks 640 accounting software companies through his research platform, and he said relatively few offer AI that handles tax effectively. His explanation is simple. Large language models are probabilistic, producing answers from patterns, while accounting is deterministic and needs the same correct answer every time. In his words, tax isn’t just arithmetic. It’s law inside a system that needs numbers to work.
Ross Martin added that much of the calculation work doesn’t need AI anyway, because rules-based software already handles calculations and decision trees well. The harder problem is context. Good tax advice depends on a client’s full circumstances, including reliefs, losses, acquisitions, retirement and life changes, and on knowing which question to ask in the first place. Drawing on her time running tax helplines, she said the hard part isn’t answering the question but getting people to ask the right one.
Both guests stressed the importance of professional judgment, especially when an accountant signs off the work. Anderson made a point every practice should remember: clients don’t pay advisers just for data entry. They pay them to take the worry of dealing with HMRC off their hands.
Meanwhile, clients are already using AI tools. Business owners type tax questions into ChatGPT, Copilot and Google’s AI answers every day. The tools may get the details wrong, but they still shape what people think they need and who they ask next.
What independent accountants need to know
AI is becoming the first stop for questions. A sole trader wondering whether they need MTD, or a director weighing up a dividend, may ask an AI assistant before they ask a person. If the answer suggests talking to an accountant, the tool may name firms, or the person will search for one straight away.
AI’s weakness is your selling point. The limits Anderson and Ross Martin describe, such as unreliable calculations, missing context and no accountability, are exactly what accountants provide. That message needs to appear in your marketing, not just in your engagement letter.
AI answers draw on what’s published. AI search tools build answers from content they can find and trust. Firms with clear, specific, well-structured pages about the services they offer and the problems they solve are more likely to be cited or recommended. Firms with a thin website and a generic services list are easier to overlook.
The questions behind the question are the opportunity. Ross Martin’s point about asking the right question is also a content strategy. If your website explains the questions clients don’t know to ask, such as the tax implications of selling a business, taking on staff or changing structure, you become the source AI tools and searchers turn to.
What forward-thinking practices are already doing
Practices that are building visibility in AI search tend to focus on four things.
They publish answers, not adverts. Each key service has its own page that answers real client questions in plain English: who it’s for, what it involves, what can go wrong and what it costs. Short, factual, quotable statements are easier for AI tools to use than marketing slogans.
They show where judgment matters. Instead of claiming to be “proactive”, they explain specific situations where professional judgment changes the outcome, such as a relief that was nearly missed, a loss that could be carried back or a restructuring that needed proper timing. Anonymised examples like these show expertise in a way generic copy can’t.
They keep their facts consistent. Firm name, address, services, specialisms and qualifications are the same on their website, Google My Business for accountants profile, directories and professional body listings. AI tools combine information from many sources, so inconsistencies weaken trust.
They update content when the rules change. A page about MTD or the latest Budget that is dated and kept current signals that the firm is active and reliable. Stale content tells both people and algorithms the opposite.
How this connects to growth
This is GEO (Generative Engine Optimisation) in practice: making sure that when AI tools answer a question about tax or accounting in your area, your firm is part of the answer. It sits alongside traditional local SEO rather than replacing it. Many of the same signals, including clear service pages, reviews, consistent business details and genuine local relevance, help with both.
The commercial logic is simple. Someone who has asked an AI tool a tax question and been told to speak to an accountant is close to buying. They already know they need help. If your firm appears at that point, with a clear explanation of what you do and an easy way to book a call, you’re likely to win the enquiry. If you don’t appear, a competitor will.
For independent accounting firms, AI SEO also levels the playing field. AI answers tend to favour relevance and clarity over big advertising budgets. A well-documented local practice with genuine specialisms can be recommended alongside, or ahead of, a national brand.
Machines can draft the answer, but people still need an accountant
Anderson’s view is that AI will improve the workflow around tax before it can be trusted with the tax itself. That gives accountants time, but not unlimited time. The firms that use it well will make sure that every AI-generated answer which ends in “speak to an accountant” leads somewhere specific, and that the place it leads is them.
Building that kind of visibility takes expertise in content, search and AI, which most practices don’t have in-house. That’s what the CharterGroup Alliance provides. It helps independent firms get found on Google, in AI answers and in their local market, backed by a network of peers working on the same challenge. If you want to be the accountant AI recommends, see what CharterGroup membership could do for your firm.
Published by the CharterGroup team
