Australian executives trust AI despite audit error findings

Workiva has published Australian findings from its 2026 Midyear Executive Benchmark Survey, showing that 21 per cent of local executives say internal audits found AI errors that reached external audiences or board members.

The survey also found that 79 per cent of executives are at least somewhat confident in the accuracy of AI output without human review, despite those audit findings.

The figures highlight a gap between executive confidence in AI and the quality controls around its use. Workiva's Australian data suggests many organisations are using AI in reporting and other workflows while still grappling with weak data quality, traceability issues and governance gaps.

Only 14 per cent of executives surveyed said their data quality was sufficient for AI use. Another 21 per cent said poor data quality had significantly blocked AI deployment in key workflows, while 65 per cent said it had at least moderately affected AI use in financial and sustainability reporting.

These findings place data management at the centre of the AI debate for finance, risk and sustainability teams. If source data is incomplete, inaccurate or difficult to verify, AI outputs may carry those same faults into reports, board papers and external communications.

Kristen Pimpini, Vice President and General Manager APJ at Workiva, said the issue starts with the information fed into AI systems.

"Bad data dressed up by good AI is still bad data; it's just faster and harder to catch. Before leaders confidently hand more decisions to AI, they need to know where the numbers came from and be able to prove it when someone asks," Pimpini said.

The survey also sheds light on executive concerns as AI becomes more widely used in Australian organisations. It found that 43 per cent of local executives are concerned about intellectual property leakage through unauthorised AI tools, while 39 per cent are concerned employees may fail to verify AI-generated data before it is used externally or presented to the board.

Audit findings

Internal AI audits appear to be exposing several weaknesses. The research found that 39 per cent of surveyed executives said their organisation's most recent AI audit uncovered data lineage gaps or traceability issues, while 29 per cent said it identified gaps in governance policies or controls.

The share reporting hallucinations or other errors that reached external audiences or board members was 21 per cent. By contrast, only 38 per cent of Australian executives said their audit validated existing controls.

That balance suggests many organisations are still in the early stages of building formal oversight around AI use. It also shows audit functions are being drawn into issues that extend beyond traditional financial controls to the management of automated decision support and generated content.

David Conley, Head of Reporting at Challenger, linked stronger data trails to the daily workload of finance teams.

"Finance teams should be spending more time supporting their businesses and its initiatives, and less time on the process of reporting and closing out the month and year end. Confidence in data, built by a clear trail of where it came from and who touched it, takes the guesswork out of reporting and limits human error from creeping in. That frees teams to focus their talents on the key business initiatives and value-added work, not ticking and tying a set of financial statements for the fifth time or manually shuffling data around," Conley said.

Use and returns

Despite the concerns, the survey indicates that Australian executives are seeing measurable returns from AI adoption. Revenue growth was the most commonly cited measure of return on investment, selected by 61 per cent of respondents, followed by time savings at 53 per cent and automation rate at 48 per cent.

Some organisations are also using more traditional finance metrics to assess impact. Internal rate of return was cited by 36 per cent, while 23 per cent said they use revenue per employee.

The reported effect on work is mixed rather than uniformly positive. While 74 per cent said AI had improved the quality of at least some work produced, 46 per cent said it had removed or automated tasks their team previously handled. At the same time, 32 per cent said it had created more tasks and review requirements.

That increase in review work may reflect the need to check outputs, confirm source data and ensure material produced with AI is fit for internal or external use. The survey's finding that 88 per cent of executives see AI training as a top priority suggests many organisations believe staff skills and oversight processes have not yet kept pace with adoption.

Infrastructure needs

The research also points to the systems executives believe they will still need as AI tools become more autonomous. More than half, 59 per cent, said they would need platforms to manage agents and automated workflows. Another 51 per cent said systems of record such as general ledgers would remain necessary, and 50 per cent said they would need software that enables traceability and audit.

The survey polled 2,272 finance, risk and sustainability professionals worldwide, including 847 C-level executives, across North America, Latin America, Europe and the Asia Pacific region, including Australia. In the Australian results, the strongest message was that confidence in AI remains high even as audits continue to uncover errors, data lineage gaps and control weaknesses.