Executives trust AI despite errors, Workiva survey finds

Workiva has published survey findings showing a gap between executive confidence in artificial intelligence output and evidence of errors. The research found that 26 per cent of surveyed executives said internal audits had detected AI mistakes before they reached external audiences or board members.

Even so, 84 per cent of executives said they remained at least somewhat confident in the accuracy of AI-generated output without human review.

That contrast sits at the centre of Workiva's 2026 Midyear Executive Benchmark Survey, which polled 2,272 finance, risk and sustainability professionals, including 847 C-level executives, across North America, Latin America, Europe and the Asia Pacific region. It also surveyed 367 institutional investors from firms in North America and the United Kingdom.

Beyond confidence levels, the data pointed to concerns about the information feeding AI systems. Only 11 per cent of executives said their data quality was sufficient for AI use, while 27 per cent said poor data quality had significantly blocked deployment in key workflows.

Reporting functions appeared especially exposed. The survey found that 71 per cent of executives said poor data quality had at least moderately affected the use of AI in financial and sustainability reporting.

Investor sentiment also featured in the findings, with 89 per cent of institutional investors surveyed saying they were concerned about the accuracy of AI-generated content in corporate disclosures.

Data concerns

The results suggest companies are trying to expand AI use while still dealing with weak data controls and inconsistent oversight. For finance teams and boards, that raises questions about how much reliance can be placed on machine-generated analysis or draft disclosures before staff review them.

Barbara Larson, Chief Financial Officer at Workiva, linked the issue to governance and verification.

"Confidence in AI without control over data quality is a liability, not a strategy. CFOs need platforms that connect AI to trusted, auditable data so every output is one they can verify and every disclosure is one they can defend," said Barbara Larson, Chief Financial Officer, Workiva.

Larson also pointed to the commercial implications of greater trust in AI systems.

"Getting this right is about more than avoiding errors. Business leaders can move faster and embed AI deeper into their operations when they trust what their systems produce. That's a real competitive edge," she said.

Infrastructure needs

The survey also examined what executives think they will need as AI tools become more embedded in financial reporting and business processes. Many pointed not to the models themselves, but to the systems around them.

Among those surveyed, 49 per cent said they would continue to need systems of record such as general ledgers. Another 45 per cent said they would need software that enables traceability and audit, while 55 per cent said they would need platforms to manage agents and automated workflows.

Those responses indicate that companies remain focused on the underlying control environment as they assess wider AI use. In regulated functions such as finance and sustainability reporting, traceability and data lineage remain central concerns because organisations may need to explain how outputs were produced and what information was used.

Jason Darby, Chief Financial Officer of Amalgamated Bank, said generic tools were unlikely to satisfy that level of scrutiny.

"Generic AI isn't enough for financial reporting. Investors, regulators, and boards expect answers they can trust. The real advantage comes from specialised AI built on governed, auditable data and paired with human judgment, giving organisations the confidence to verify what AI produces and stand behind the decisions and disclosures that follow," said Jason Darby, Chief Financial Officer, Amalgamated Bank.

The findings come as companies face growing pressure to show that AI systems used in reporting, risk and operational decision-making can be reviewed and challenged. The survey suggests many executives remain positive about the technology's output, but fewer believe their underlying data is ready for broader use.

That mismatch may become more visible as AI-generated material appears in investor communications, board papers and formal disclosures, where errors can carry wider consequences than in internal experiments. Workiva's figures show that concerns about AI accuracy are shared not only inside companies, but also by the investors who rely on those disclosures.