Only 34% trust AI agents despite 86% deploying them

Boomi says the enterprise trust gap around agentic AI has less to do with model capability and more to do with weak data foundations.

According to new Boomi research, 86% of enterprises have moved from agentic AI pilots into production-level deployment, but only 34% say they trust the actions their agents are taking. The company argues that the problem is not model intelligence, but data quality, integrations, governance, and controls.

Boomi split respondents into readiness groups, with the bottom quartile labeled “agentic chaos.” Even there, 77% of businesses have still deployed agents, or are moving them into production. Boomi warns that pressing ahead without the right foundations could expose companies to compliance fines, lost customers, and operational downtime.

Recommended reading

UK data centre boom hits a water reality check

At the other end, the top quartile was categorized as “agentic control,” and 55% of those organizations said they are highly confident in their agents' decisions and actions.

The report also suggests many companies are solving for the wrong problem. Among businesses in agentic chaos, about 51% are prioritizing AI model or tool maturity instead of addressing the factors Boomi says actually determine trust.

“Agents can only be trusted to act on data that’s been properly activated, connected, and governed, and most companies deployed agents before they did that work.”

The findings point to a familiar enterprise pattern: pressure to show AI value and ROI is speeding up deployment before the underlying systems are ready, leaving organizations to fix data and governance issues after agents are already in production.