AI agents are becoming mainstream in the enterprise, but most organizations still don’t trust them, according to new research commissioned by Boomi.

The Forrester Consulting study, which surveyed 409 senior IT and technology decision-makers across North America, Europe, and Asia-Pacific, found that 86 percent of organizations have moved beyond AI agent pilots. Yet only 34 percent trust the decisions their AI agents make.

The report found that organizations in what it calls “agentic chaos” are deploying AI despite weak governance and integration, exposing themselves to an average of $2.1 million in additional costs from compliance penalties, operational disruptions, customer losses and rework.

By contrast, organizations classified as having “agentic control” reported much higher confidence, with 55 percent trusting their AI agents, compared with just 22 percent among organizations in agentic chaos.

“The trust problem with agentic AI is really a data problem,” said Steve Lucas, chairman and CEO of Boomi. He said AI agents can only be trusted when they are built on connected, well-governed data.

The study identified integration, rather than more advanced AI models, as the biggest factor separating organizations that trust their AI agents from those that do not.

Nearly half (46 percent) of organizations with agentic control use integration platform as a service (iPaaS) to power AI workflows, compared with 25 percent of organizations in agentic chaos.

Companies with stronger AI governance also reported better results, including productivity gains (59 percent), increased innovation (51 percent), reusable business capabilities (46 percent), and automation of repetitive tasks (45 percent).