Widening 'AI proof gap' in firms

A new Grant Thornton survey found 78% of senior business leaders are not strongly confident their company could pass an independent artificial intelligence governance audit within 90 days. (grantthornton.com) Grant Thornton released the findings on April 13, 2026, after surveying 950 C-suite and senior business leaders in the United States in early 2026. The firm said only 12% believe their workforce is truly ready for artificial intelligence. (grantthornton.com) An artificial intelligence governance audit is a test of whether a company can show how its systems work, who owns the decisions, and what happens when something goes wrong. Grant Thornton said many companies are deploying tools they cannot fully explain, measure, or defend. (grantthornton.com) The survey points to a split between spending and oversight. Grant Thornton said three in four boards have approved major artificial intelligence investments, but only 52% have set clear governance expectations and 54% have folded artificial intelligence risk and opportunity into regular board or committee oversight. (grantthornton.com) Operations executives said the strain is showing up in day-to-day performance. Half of operations leaders told Grant Thornton they need a formal artificial intelligence strategy or governance plan within six months, and 46% of leaders said controls and compliance failures are a leading reason artificial intelligence underperforms. (grantthornton.com) Grant Thornton tied that governance gap to business results. Its report said companies with fully integrated artificial intelligence were nearly four times as likely to report artificial intelligence-driven revenue growth as companies still in pilot mode, 58% to 15%. (grantthornton.com) The pressure to document and control artificial intelligence systems has been building outside one survey. The National Institute of Standards and Technology published its Generative Artificial Intelligence Profile in July 2024 as a companion to its Artificial Intelligence Risk Management Framework, giving companies a federal reference point for managing model risks. (nist.gov) Europe also moved from principle to enforcement timelines. The European Commission said the European Union’s Artificial Intelligence Act entered into force on August 1, 2024, starting a phased rollout of obligations for companies that build or deploy covered systems in the bloc. (commission.europa.eu) Other researchers are finding a similar gap between policy and practice. Thomson Reuters Foundation said in February 2026 that 48% of 1,000 companies it studied disclosed artificial intelligence strategies or guidelines, but only 38% of United States companies had published artificial intelligence policies. (thomsonreuters.com) The closing argument in Grant Thornton’s survey was not that companies should slow artificial intelligence spending. It was that the firms moving fastest with the technology are the ones that can already show the receipts. (grantthornton.com)