Many firms fail AI audits
A recent survey says lots of companies can't pass an AI-governance audit within 90 days, exposing a gap between board-level enthusiasm and operational readiness. The finding was described as a widening “AI proof gap” in a report cited by Insurance Business, which highlights governance shortfalls firms would need to remedy to withstand review. (insurancebusinessmag.com)
Most companies using artificial intelligence still do not believe they could pass an independent AI-governance audit within 90 days. (grantthornton.com) Grant Thornton said its AI Impact Survey, conducted in early 2026, drew responses from nearly 1,000 senior business leaders in the United States across industries. The firm said 78% lacked full confidence they could clear that audit window. (grantthornton.com) The same survey found only 12% of leaders said their workforce is “truly AI-ready,” even as companies keep spending on tools and pilots. Grant Thornton called the disconnect an “AI proof gap” between investment and accountability. (businesswire.com) An AI-governance audit is a document check and control check: a reviewer asks what systems a company uses, what data they touch, who approves them, how risks are tested, and how problems are logged. The National Institute of Standards and Technology says AI risk management should cover the full lifecycle, not just the launch. (nist.gov) That matters in 2026 because formal rules are moving closer to day-to-day operations. The European Union’s Artificial Intelligence Act entered into force on August 1, 2024, and obligations for deployers of high-risk systems are set to apply from August 2, 2026. (commission.europa.eu) (artificialintelligenceact.eu) The European Union law is built around “safe and trustworthy” systems, with duties tied to risk level rather than marketing claims. That means companies selling or using artificial intelligence across borders can face questions about records, oversight, transparency and fundamental-rights safeguards. (eur-lex.europa.eu) Grant Thornton also said companies with fully integrated artificial intelligence were nearly four times as likely to report revenue growth as companies still stuck in pilot programs. The firm’s argument is that stronger governance is not separate from adoption; it is part of scaling it. (businesswire.com) Insurance Business, which first highlighted the survey for that audience on April 13, said the shortfalls leave firms exposed when boards, regulators, customers or insurers ask for evidence that systems are safe and effective. The publication framed the problem as weak operating discipline behind board-level enthusiasm. (insurancebusinessmag.com) The gap is not only about regulation. The Organisation for Economic Co-operation and Development says its AI principles are meant to push systems that respect human rights, transparency and accountability, which is the same evidence trail an audit usually demands. (oecd.org) So the immediate test for companies is less about whether they have bought artificial intelligence software and more about whether they can show who is responsible for it, what it does, and how it is controlled. The survey suggests most still cannot do that on a 90-day clock. (grantthornton.com)
