Businesses are increasingly testing the waters with AI agents, but trust issues persist.
A survey by integration platform Workato and Harvard Business Review of more than 600 tech leaders found that most respondents find agents more trustworthy outside of core business areas, while just 6% said they fully trust agents with essential end-to-end business processes. But that hasn’t curbed interest: 86% said their companies plan to invest more in agentic AI over the next two years.
More than two in five (43%) of respondents said they only trust agents with routine operational tasks, 39% said they delegate agents to “supervised use cases or complex, noncore processes,” and 8% said they don’t trust agents with business operations at all.
“This report is almost exactly the conversation I had with [CEO friends],” Workato Chief Information Officer Carter Busse told us. “When I read it, I’m like, ‘Oh my gosh, this actually is real data [for] the conversations I’m having where, ‘Yeah, we bought all these Chats and Claudes and Geminis and people can summarize their emails and look up stuff in their calendar and help me write a nice letter.’ But is that real work? It’s not.”
As AI systems that purport to perform tasks autonomously have sashayed into vogue in the corporate world, trusting these systems to handle important business functions has become a sticking point. That’s especially true as agents are integrated into various workplace platforms through protocols.
Busse said many agents currently in use might perform simple tasks like creating IT tickets. But those that can handle more complicated processes are still in their earliest stages of adoption, he said.
“We’re very, very early,” Busse said. “When I think of agents, I think of real work…A real work agent does a multistep, complex process that’s trusted. That’s going to take a long time to get there.”
The report also found that agents do not yet meet expectations across all…