About this article This article is personal, general commentary. It does not represent or carry the endorsement of any employer, ICAEW, AAT or another professional body; it is not accounting, tax, legal, investment, employment or data-protection advice; and it creates no adviser–client relationship. No client or employer information has been used. Verify current rules and obtain advice appropriate to your circumstances before acting.
Artificial intelligence is no longer a distant prospect for the accounting profession. It is already present in bank-feed matching, invoice capture, anomaly detection, forecasting, spreadsheet review and the drafting of routine communications. In my experience, the best results come when we treat AI as an assistant: useful for speed and pattern recognition, but never the person accountable for the figures.
What do we mean by AI accounting?
The label covers several different technologies. Some are rule-based and predictable; others learn patterns from data; generative AI produces text or summaries from prompts. That distinction matters. A bank rule that follows an agreed instruction is not the same as a chatbot that can produce a plausible but incorrect answer.
For a small business, the most useful applications are usually modest: extracting details from invoices, suggesting transaction categories, matching receipts to payments, highlighting unusual entries, preparing a first draft of a cash-flow commentary, or turning an accountant’s notes into a clearer client email.
ACCA’s Alistair Brisbourne puts the professional challenge neatly: “new dynamics are being introduced to the traditional trust mechanisms that underpin the accountancy profession.”
The advantages — and the boundary
Used well, AI can shorten repetitive work, improve consistency and bring exceptions to the accountant’s attention earlier. That creates more time to ask the questions that matter: Why has margin changed? Is cash tight because customers are paying slowly? Is the forecast based on realistic assumptions?
The boundary is accountability. Software does not know the full commercial context, cannot accept professional responsibility and may not explain how it reached an answer. I would not allow an AI-generated journal, tax conclusion, forecast or client recommendation to pass into final use without a competent person checking the source data, assumptions and output.
A sensible starting point
- Choose one low-risk, repetitive process rather than attempting a wholesale transformation.
- Use approved business tools and do not paste identifiable client data into a public AI service.
- Define what the human reviewer must check and retain evidence of that review.
- Measure time saved, corrections required and whether the output genuinely improves client service.
- Stop or redesign the process if errors are difficult to detect or explain.
AAT’s practical reporting shows accountants using AI to support reconciliation, identify discrepancies and improve communications. That is a realistic picture: incremental gains rather than an autonomous finance department.
My conclusion
AI will become part of ordinary accounting software, much as cloud bookkeeping did. The firms and businesses that benefit most will not be those that automate everything first. They will be those that combine clean data, well-designed controls and professional scepticism with technology that solves a defined problem. The accountant’s role remains central: to test the evidence, explain the numbers and take responsibility for the advice.
Continue reading on The Perry
- AI Bookkeeping Automation: Saving Time Without Losing Control
- From Bookkeeping to Better Decisions: AI Forecasting and Advisory Work
- Making Tax Digital for Income Tax in 2026: a practical first-year checklist
References and further reading
- ACCA, “AI Monitor: Exploring the trends, innovations and challenges of artificial intelligence” (accessed 29 August 2026).
- AAT Comment, “How are accountants and bookkeepers using AI in their day-to-day practice?” (19 August 2026) (accessed 29 August 2026).
- ICAEW, “Generative AI and ethics” (accessed 29 August 2026).
- ICAEW, “Global research on AI and the future of the profession” (17 April 2025) (accessed 29 August 2026).