Generative AI is increasingly being used in professional services to support research, document review, drafting, summarisation and issue identification. For accounting firms, the opportunity lies in using AI to improve efficiency in audit, tax, compliance and advisory work, particularly where financial reporting issues overlap with contractual, regulatory or legal considerations.
However, AI should be treated as a working tool, not as a final authority. Where legal documents or legal concepts are involved, accountants should apply disciplined professional judgement and avoid presenting AI-assisted analysis as legal advice.
Key Impacts
Audit and Risk Assessment
AI can assist audit teams by helping to organise large volumes of information and identify potential risk areas from contracts, board papers, policies, management reports or correspondence. This may support the audit planning process by highlighting matters that require further review, such as revenue recognition, lease terms, debt covenants, contingent liabilities, related-party indicators, impairment triggers or disclosure obligations.
The benefit is not that AI replaces audit judgement, but that it can help teams work through complex information more efficiently. Auditors must still verify the output, assess relevance, and apply professional scepticism.
Agreement Review
Accounting firms often review agreements to understand financial reporting, tax or business implications. AI can assist by summarising key terms and flagging clauses that may affect accounting treatment.
For example, contract terms relating to termination, renewal, variable consideration, penalties, warranties, indemnities, financing obligations or tax gross-ups may affect accounting estimates, provisions, disclosures or audit procedures.
The review should be clearly limited to accounting, audit, tax or business implications. Questions about enforceability, legal rights, remedies or interpretation of law should be referred to qualified legal counsel.
Tax and Regulatory Compliance
AI can support tax and compliance teams by summarising legislation, guidance, client facts and transaction documents. It can help prepare issue lists, identify missing information and organise technical considerations across jurisdictions.
However, tax and regulatory work often involves legal interpretation. AI-generated summaries should therefore be checked against authoritative legislation, official guidance, firm technical materials and specialist review.
Client Advisory and Business Operations
AI can assist in preparing first drafts of client briefings, board papers, management questions and advisory notes. It may also help translate technical issues into practical business implications.
For client-facing work, firms should clearly distinguish between accounting or tax observations and legal advice. Where a matter involves legal enforceability, statutory rights, litigation risk or regulatory interpretation, the firm should recommend that the client obtain legal advice.
Practical Issues
Quality of Inputs
The usefulness of AI output depends heavily on the quality of the instructions given. Vague requests can produce broad or unreliable answers. Teams should provide sufficient background, identify the relevant jurisdiction, define the purpose of the review, state the desired output, and specify whether the focus is audit, tax, compliance or business reporting.
Risk of Unauthorised Legal Advice
Accountants may need to review legal documents, but that does not mean they are providing legal advice. Firms should ensure that AI-assisted work does not cross into legal conclusions unless the firm is authorised and qualified to provide such advice.
Where appropriate, client communications should state that the review is limited to accounting, audit, tax or commercial reporting implications and does not constitute legal advice.
Hallucination and Verification
AI tools may generate incorrect or unsupported information, including inaccurate legal references, fabricated citations or overconfident conclusions. This is especially important where legal, tax or regulatory sources are involved.
Before relying on AI output, teams should verify cited materials, read source documents directly, check whether the analysis is logical and compare conclusions against firm-approved technical guidance.
Confidentiality and Data Security
Client agreements, audit files and tax documents often contain confidential information. Firms should only use AI tools that comply with internal policies, client confidentiality obligations and data protection requirements.
Before uploading client information, firms should consider whether the AI platform stores prompts, uses data for model training, permits third-party access, or allows deletion of conversation history.
Documentation and Audit Trail
Where AI is used in audit or advisory work, teams should retain appropriate documentation. This may include the purpose of using AI, the documents reviewed, the AI output, the professional review performed, source verification and final conclusions.
AI-generated content should not be treated as audit evidence unless independently verified against reliable source documentation.
Action Points
Accounting firms should consider adopting a controlled approach to AI use by:
- approving specific AI tools for client work;
- training staff on responsible AI use;
- creating internal guidance for audit, tax and advisory use cases;
- requiring verification of AI-generated outputs;
- maintaining documentation where AI assists professional work;
- including suitable disclaimers where legal issues are discussed;
- escalating legal interpretation to qualified legal counsel; and
- reviewing engagement terms to clarify the scope and limits of AI-assisted work.=
AI can improve efficiency in audit, tax, compliance and advisory work by helping professionals organise information, identify issues and prepare first drafts. Its value is strongest when used within a structured professional framework.
For accounting firms, the key is responsible use: protect confidentiality, verify outputs, document review procedures and clearly separate accounting or tax analysis from legal advice. AI should support professional judgement, not replace it.