December 09, 2025
Enterprises Are Pouring Money Into AI They Don’t Trust
Big AI budgets are becoming the norm across enterprise tech, even as confidence in autonomous systems remains hard to find.
New research from Workato surveyed more than 600 technology decision-makers and found that only 6% currently trust agentic AI to handle core business processes on its own, yet spending continues to accelerate.
Eighty-six percent of respondents expect their organisation’s investment in agentic AI to rise over the next two years, suggesting companies are backing the technology long before fully believing in it.
Solving Fundamental Problems First
The tension is altering the pace and scope of AI rollout inside enterprise teams. Instead of handing over control, organisations are trying to wrap agentic systems in layers of structure and supervision. Seventy-four percent are either already implementing or planning enterprise orchestration platforms, a software designed to connect data sources, track automated activity, and limit how far AI can operate independently.
The priorities behind these investments are practical rather than visionary. Eighty-two percent say improving application connectivity matters most, ensuring that systems can exchange information smoothly. Eighty percent focus on providing AI agents with accurate business context and real-time data, an attempt to prevent decision-making from drifting into guesswork. Over the next two years, 76% expect to continue spending on platforms that allow agents to plug into core workflows without removing human oversight entirely.
Carter Busse, Chief Information Officer, Workato, said: “While the majority of businesses are doubling down on AI investments, IT leaders do not trust AI agents to make those transformation projects successful. These results are consistent with what I hear from CIOs every day who are looking for an enterprise-ready approach to deploy AI that blends orchestration, governance, and enterprise-grade skills so that it acts predictably and securely in mission-critical systems. The companies that solve this fundamental problem will be the ones who win in this new era of AI.”
Agentic AI Impact on CX
The findings reveal customer experience leaders are cautious when it comes to expectations around agentic AI. The strongest projected benefits of agentic AI centre on operational performance rather than customer-facing change. Improved efficiency and productivity top the list at 67%, while 62% anticipate cost reductions. Employee agility and responsiveness follow at 54%, with wider benefits that may eventually translate into smoother service delivery.
Again, most organisations expect AI to help their operations before it improves customer experience. Just under half of them (48%) believe agentic AI will materially improve customer experience, and 46% expect stronger decision-making as a result of adoption.
Companies plan to deploy AI agents first in IT (56%) and operations (55%), targeting workflows, system coordination, and internal services. Marketing (32%) and sales (31%) trail behind, areas where failures are visible to customers and harder to contain.
The approach remains conservative: test, assist, watch, but don’t surrender control. After all, the AI race won’t be won by flashy tools, but by knowing when humans still need to stay in the loop.
