AI in External Audit: Opportunity or Threat?

Large language models and machine learning algorithms are no longer pilot projects inside major audit firms — they are being deployed in live engagements for tasks ranging from automated journal entry testing to natural language review of contracts and Board minutes. The speed gains are substantial: what formerly required a team of seniors three weeks of sampling can now be completed across an entire transaction population in hours. However, the shift from sampling to full-population testing introduces its own interpretive challenges, and the audit profession has not yet established consensus on how AI-assisted procedures satisfy the evidence requirements of AS 1105 or ISA 500.

The PCAOB's December 2025 staff guidance on the use of technology-assisted analysis in attest engagements represents the most substantive regulatory intervention to date. The guidance confirms that auditors remain responsible for the design, execution, and evaluation of technology-assisted procedures and cannot delegate professional scepticism to an algorithm. Crucially, it requires auditors to understand the inputs, logic, and limitations of AI tools used in the engagement at a level sufficient to evaluate the reliability of their outputs — a standard that many firms are struggling to operationalise given the opacity of large language model reasoning.

Audit committees occupy a critical oversight position in this transition. They should be asking their audit firms directly: which AI tools were used in the current engagement, what validation testing has been performed on those tools, and how errors or hallucinations in AI output are detected and corrected before conclusions are drawn. Firms that cannot answer these questions with specificity are not yet operating to the standard that the PCAOB expects, and audit committee members should treat that uncertainty as a quality risk.

From a broader competitive perspective, AI adoption in audit is accelerating the consolidation of market share toward larger firms with the capital to invest in proprietary AI infrastructure. Mid-sized firms face a difficult choice: invest heavily in technology partnerships or risk falling behind on both quality and efficiency. INNOVATIVE NATIONAL TAX & UPKEEP INTERNATIONAL TALLY PTY LTD's Audit Quality team works with firms at all points of this transition, helping design AI governance frameworks that satisfy regulatory expectations while realising the genuine efficiency gains that make adoption worthwhile.