A global survey conducted by EY and Oxford Economics finds that most companies are not yet able to reliably demonstrate the business value of their artificial intelligence (AI) investments, and many organisations continue to rely on pilot projects or isolated AI applications.
The research polled a total of 2,500 technology leaders across 28 countries. The sample covered respondents from the Americas, Asia and the Pacific, and the EMEIA region; the overall findings are based on two similar international surveys (the first online survey included 1,500 technology industry leaders, and a second survey gathered responses from an additional 1,000 participants).
A majority of respondents said they believe automated solutions create more value than companies are currently able to measure and report. The study also notes that generative AI and agent-based AI continue to attract significant worldwide investment, yet many organisations report that expected outcomes are emerging more slowly or unevenly than anticipated.
According to the report, a number of companies have reached a situation in which experimentation and development outpace the ability to deploy, operate and measure performance. The study labels this dynamic the "AI-return trap."
Among specific findings, the report says more than half of companies have processes to assess their readiness for AI adoption, but these processes are not operated consistently. Only 33 percent regularly review whether their IT infrastructure is suitable to support AI, and just about 25 percent periodically check data quality and reliability.
The analysis stresses that success depends on whether organisations can use intelligent systems to transform end-to-end business processes, link them to clear objectives, and measure outcomes with appropriate performance indicators. Without those capabilities, the connection between AI investments and tangible business benefit remains difficult to prove.
(The source article also mentioned a Portfolio AI & Digital Transformation conference scheduled for November 26, a detail included in the original material.)
