Audit firms embed AI while tightening validation controls

Audit and accounting firms are moving beyond trial projects with artificial intelligence. New research points to broad adoption, alongside clear expectations for human oversight and accountability.

A global survey of more than 1,000 audit and accounting professionals by IDC, sponsored by Caseware, found that 66% of respondents said AI was embedded in their firm's strategy, widely used in selected functions, or in pilot projects. More than half (53%) agreed or strongly agreed that AI tools can improve audit quality.

The results suggest AI is becoming part of firms' core operating models, but caution remains when the technology could influence audit conclusions. Respondents reported strong expectations for validation, governance, and training as AI moves from experimentation into production workflows.

Validation expectations

Most respondents expect AI to reshape the sector in the coming years. The study found that 76% believed AI would fundamentally transform audit and accounting over the next decade.

At the same time, the survey highlighted a strong professional norm around responsibility for judgments informed by AI outputs. Nearly two-thirds (64%) said validation should always be required when relying on AI in professional judgments used to support audit conclusions.

This reflects long-standing requirements in audit and assurance around evidence, documentation, and review. It also suggests that firms adopting AI at scale face pressure to define clear controls over model outputs, data inputs, and the circumstances in which staff can rely on automated analysis.

David Marquis, Chief Executive Officer of Caseware, said the findings show broad acceptance of AI alongside clear guardrails.

"What stands out in this research is not hesitation about AI, but clarity," said David Marquis, Chief Executive Officer, Caseware. "Audit and accounting professionals recognize the value AI can bring, while setting firm expectations around validation, judgment and accountability. Trust remains the foundation of the profession. At Caseware, we're committed to developing technology that enables the profession to evolve while safeguarding that trust. Our position as an established industry authority is built on consistently delivering intelligence across the workflow through innovation the profession can depend on; innovation that maintains the rigorous standards that audit and accounting experts demand."

In audit and accounting, AI tools are increasingly used for document review, risk assessment, anomaly detection, and drafting support. The research suggests firms are balancing those uses against questions of reliability and readiness, particularly when automated outputs might affect material conclusions.

Integration focus

IDC said AI's impact will depend on how well firms integrate tools into existing processes, rather than deploying stand-alone solutions. That includes aligning AI systems with established methodologies, quality controls, and governance frameworks.

"As AI becomes embedded into day-to-day workflows, its value is increasingly defined by how well firms integrate these tools into established methodologies, quality controls and governance frameworks. Firms that prioritise this thoughtful integration will be better positioned to maintain the trust that defines the profession," said Mickey North Rizza, Group Vice-President, Enterprise Software, IDC.

The emphasis on integration reflects the structure of audit work, where engagement teams follow standardised programs, apply defined thresholds, and manage evidence through review hierarchies. It also reflects regulatory expectations around independence, professional scepticism, and accountability requirements that do not transfer to automated systems.

Adoption barriers

The survey also asked respondents to identify the single biggest barrier to AI adoption in their practice. Cost of implementation ranked first, cited by 34% of respondents, followed by a lack of technical talent (30%).

These findings highlight operational hurdles that can slow deployment even as interest grows. Implementation costs can include software licenses, data preparation, process redesign, security work, and change management. Skills shortages can limit both the ability to deploy tools and the ability to challenge outputs, reinforcing calls for mandatory validation and clearer governance.

Respondents included senior decision-makers across a range of firm sizes. About a quarter worked in organisations with 500 or more employees, and a similar share worked in organisations with 11 to 50 employees. The survey covered multiple geographies, with the largest share based in the US, followed by the UK and Canada, then Australia, Germany, and the Netherlands.

Caseware, which supplies audit and accounting software, sponsored the study as part of its broader focus on AI-enabled workflows for assurance and accounting. It plans to release additional survey findings in future updates.