A significant governance gap has emerged in UK accountancy: over 40% of accounting professionals are now losing up to 10 hours per week correcting erroneous tax and expense advice generated by generic chatbots lacking specialised financial knowledge. According to Business Rescue Experts' April 2026 survey, this 'AI slop'—low-quality or hallucinated output from large language models untrained on accounting-specific rules and regulatory obligations—is creating material productivity drag and, more critically, introducing compliance risk. Firms deploying ChatGPT, Claude or similar general-purpose AI tools for tax advice without domain-specific guardrails are inadvertently outsourcing professional judgment to systems incapable of reasoning about regulatory context.

The FRC's response signals a fundamental shift in supervisory focus. Rather than issuing prescriptive rules about which AI tools are permitted, the Financial Reporting Council has moved toward preventative supervision that scrutinises firm culture and technological infrastructure. This approach reflects recognition that AI governance cannot be bolt-on compliance; it must be embedded in how firms evaluate, validate and monitor tools before deployment. For accountancy practices, this means the FRC is likely to probe during inspections whether firms have processes for stress-testing AI outputs against tax law, identifying conflict-of-interest scenarios, and documenting professional accountability when AI is used in client-facing advice.

The regulatory risk extends beyond FRC oversight into AML and tax compliance obligations. Under Money Laundering Regulations (MLR 2017) and benefit-of-the-doubt requirements, accountants advising on tax structure or beneficial ownership cannot delegate that analysis to unvalidated AI systems. If an accountant relies on a chatbot's tax opinion that contradicts HMRC guidance or misses beneficial ownership red flags, liability falls squarely on the firm, not the vendor. Similarly, SRA rules for legal advisors working in accountancy contexts require that professional judgment remain human-driven, particularly on matters where regulatory consequence is material.

The FRC's shift toward culture and infrastructure scrutiny creates immediate governance obligations for UK accountancy firms. Organisations must now document their AI tool evaluation processes, maintain audit trails of AI usage in client advice, and establish quality gates where senior practitioners review AI-generated output before client delivery. Generic AI tools are not inherently prohibited, but their use without specialised validation and oversight governance creates regulatory vulnerability. Trovix relevance: Trovix Brief delivers accountancy firms rapid access to FRC guidance updates, AML/tax compliance rule changes, and sector-specific AI governance best practices—enabling partners to validate AI tool outputs against current regulatory obligations and build institutional knowledge of which domain-specific AI systems align with FRC preventative supervision expectations.