The company Workiva has presented the results of its 2026 Executive Reference Survey (2026 Midyear Executive Benchmark Survey). It reveals that, although 84% of managers surveyed say they have at least moderate confidence in the accuracy of AI results without human review, one in four (26%) say that internal audits have detected AI errors that reached external audiences or board members. This data from Workiva suggests that there is a disconnect between what managers believe and what the evidence demonstrates.
As AI transforms businesses at unprecedented speed, a new risk is emerging: that the transformation is moving faster than the ability of the managers and leaders responsible for overseeing it to anticipate its implications.
“Reliance on AI without control over data quality is a risk, not a strategy. CFOs need platforms that connect AI with trusted, auditable data, so every result can be verified and every information disclosed can be defended,” said Barbara Larson, CFO at Workiva. “Doing it right goes far beyond avoiding mistakes. Business leaders can move faster and integrate AI more deeply into their operations when they trust what their systems generate. This is a true competitive advantage.”
Data quality threatens AI accuracy
Leaders largely agree that poor data quality is limiting the use of AI in their organizations, suggesting that enterprise AI can only scale when there is trust in and control over the source data. Institutional investors express similar concerns regarding the accuracy of AI-generated content.
● Only 11% of managers consider that the quality of their data is sufficient for the use of AI.
● 27% say poor data quality has significantly hampered the deployment of AI in key processes.
● 71% say poor data quality has at least moderately affected the use of AI in financial and sustainability reporting.
● 89% of investors are concerned about the accuracy of AI in disclosed corporate information.
Over time, poor or unverifiable data can erode credibility with stakeholders, and organizations that fail to address data quality and investor concerns risk falling behind more proactive competitors.
As the use of AI grows, managers demand an infrastructure to support it
For financial reporting teams, and especially for CFOs, the tools that allow you to govern and control the results of AI are as important as the AI itself. As AI agents become more autonomous, leaders believe the need for specialized infrastructure will increase, and they point to several priorities¹:
● 49% say they will continue to need record-keeping systems, such as accounting ledgers.
● 45% indicate that they will need software that allows traceability and auditing.
● 55% say they will need platforms to manage AI agents and automated workflows.
“Generic AI is not enough for financial reporting. Investors, regulators and boards of directors expect answers they can trust. The real advantage lies in specialized AI, based on governed and auditable data, and combined with human judgment. This gives organizations the confidence to verify what AI produces and support the decisions and disclosures that come from it,” said Jason Darby, chief financial officer (CFO) of Amalgamated Bank.
The real advantage lies in specialized AI, based on governed and auditable data, and combined with human judgment
The Workiva 2026 Midyear Executive Benchmark Survey surveyed 2,272 finance, risk and sustainability professionals, including 847 C-level executives, from organizations in North America, Latin America, Europe and the Asia-Pacific region. The survey also included 367 institutional investors from firms in North America and the United Kingdom.
