A survey of about 500 companies, run by Dresner Advisory Services, found something worth paying attention to if you work in a business that uses technology: the job of making AI actually pay off is landing on the desk of the CIO or CTO, not the data chief.
In more than half of the companies surveyed, the chief information officer or chief technology officer is now formally responsible for AI. Another quarter put that job with a business unit leader instead. The chief data officer, who used to run the show on data and analytics, is losing ground.
This matters because it tells you where the accountability sits when AI spending gets questioned by the board. If your company has a CIO or CTO, that person is now the one who has to show results, not just keep the lights on.
The survey also reports strong financial returns. More than half of respondents say they have gotten back more than 12 percent on their AI spending, and about 1 in 10 say returns topped 50 percent. That sounds great, but it should be read carefully.
A separate and widely cited study from MIT this year found that 95 percent of generative AI pilot projects at companies failed to show any measurable financial return at all. That is a massive gap between two pieces of research from the same year, and it is worth knowing the article behind these numbers was written by a consultant who works for the firm that ran the survey. Surveys run by advisory firms tend to attract companies already further along with technology, so take the rosy numbers here as one data point, not the full picture.
A more reliable finding in the same research: companies that had already succeeded with basic business intelligence and reporting tools before AI arrived are the ones getting the best AI results now. If your company never got its regular data and reporting in order, AI will not fix that for you. It will likely just repeat the same mess faster.
The survey also found that more than half of companies now build their own AI tools instead of buying them from software vendors. That is a real signal, but mostly for larger companies with in-house technical teams. For most smaller and mid-sized businesses, buying AI features already built into the software you use is still the far cheaper and safer route.
One more trend to watch: companies are still cautious about letting AI systems act on their own without a person checking the work first. Right now, under 1 in 10 companies let AI operate fully on its own in specific areas. But most expect that to flip within two years, with more than half planning to let AI act autonomously by then.
There is also a separate trend the survey does not mention: many companies are now creating a brand new role, a Chief AI Officer, to sit above or alongside the CIO. That role barely existed a year ago and is now common at large companies. So while this survey says CIOs and CTOs are taking charge, do not be surprised if the next survey says a different, newer title has taken over instead.
For any business leader, the takeaway is simple. Know who owns AI accountability in your company today, question any AI return figures you are given, and fix your basic data habits before expecting AI to deliver anything lasting.