Category: World-ai
Companies keep expanding what they let AI do. Fewer of them can say with confidence that it's actually working.
A survey of more than 100 C-suite and technology executives across the U.S. and Canada, published Monday by HFS Research and TCS, found that about a third of leaders say AI consistently delivers the outcomes it's supposed to, earns regulator confidence and stays under control. The rest aren't so sure.
The Trust Gap Is Wider Than the Deployment Gap
Roughly a quarter of executives surveyed said they have the governance and controls in place to scale AI across their organization. Only about one in six said they'd trust autonomous AI with critical work.
That's a notable spread. Most companies have moved well past pilot programs. Fewer of them have built the oversight structures to match.
"We've spent the last few years asking whether we can begin to deploy these tools," said Dana Daher, executive research leader at HFS Research, in comments to CIO Dive. The harder question now, she said, is whether businesses can actually depend on the results.
Adoption Without Payoff
The reliability question compounds an existing problem. A separate SAP report published in July found that AI is helping employees surface insights and engage customers, but isn't consistently saving companies time or money once inconsistent pricing models and vendor fragmentation get factored in.
Employee readiness isn't helping either. A Forrester survey found that only 16% of information workers say they highly understand the AI tools they're using.
Put together, the picture is a familiar one in enterprise technology: spending accelerates well before the workforce and the governance catch up to it.
That gap tends to get expensive. Retraining a workforce or retrofitting a governance program after AI is already embedded in daily operations costs considerably more than building both in from the start.
What Reliability Actually Means
The HFS report doesn't define reliable AI as flawless AI. It defines it as AI whose failures the enterprise can detect, explain and correct quickly.
That distinction matters for how companies build oversight. Daher argues the strongest signal of trustworthy AI isn't a technical benchmark. It's a clear human-in-the-loop structure with defined accountability.
"You need to train the person, figure out what they're accountable for," Daher said.
Right now, most companies aren't measuring what actually matters. The report found technical metrics like uptime and accuracy get tracked far more often than whether the AI delivered the business outcome it was deployed for.
There's a governance dimension too. As AI systems get more complex, it becomes harder for executives to explain how a given decision was reached.
"There's this really big tension around trust that's happening because we're being asked to adopt AI in every part of our work and lives and we don't understand how the responses are coming about," Daher said.
What This Means for Miami
South Florida's AI adoption has leaned heavily on regulated, relationship-driven industries: finance, legal services, healthcare and real estate. Those are exactly the sectors where the gap between deploying AI and trusting it carries the most consequence.
A law firm or wealth management shop rolling out AI tools doesn't just need the technology to work. It needs to be able to explain, to a regulator or a client, why a given output was correct and who's accountable if it wasn't.
Miami's enterprise AI buyers may be better served asking vendors how failures get caught and explained than asking how capable a model is in a demo. The HFS findings suggest most companies still can't answer that question about their own deployments, and the businesses that figure it out first will have a real edge over competitors still chasing capability instead of reliability.
That's a different kind of AI conversation than the one most vendor pitches are built for, and it may be the one South Florida's regulated industries need to start having.