Denodo's new research AI Trust Gap Report has revealed that agentic AI is facing a critical trust crisis.
The study revealed that 63% of organisations said that identifying the most relevant and trustworthy data, or preparing it for consumption, is a significant barrier to AI deployment. Moreover, 66% of respondents insist that AI data must be accessed in near real-time to be considered trustworthy.
"AI is rapidly shifting from systems that merely answer questions to systems that take autonomous action, and this transition changes the data requirement entirely," said Richard Jones, vice president and general manager for Asia Pacific and Japan at Denodo. "When an AI agent triggers a business outcome, there is zero room for stale or ungoverned data.
Technical hurdles
Based on a global study conducted by Arlington Research, the research highlights that technical hurdles hinder trust towards agentic AI.
The majority (67%) struggle with AI data security and access controls, which are vital requirements for safe agentic operations.
Regarding scale and complexity, over 40% 42% of organisations say they pull from over 400 original data sources for their AI initiatives.
Moreover,nearly 60% of respondents report difficulty optimising performance for the intensive workloads required by large-scale AI.
In one crucial aspect, Singapore is unique. Globally, the public sector lags in modern data management (just 52% adopting lakehouse architectures vs. 66% in other industries). On the other hand, the public sector leads in technology adoption in Singapore, driven by long-term investment in digital capability.
Jones added: "To scale agentic AI with confidence, businesses must move beyond static data silos and adopt a foundation of live, governed, and contextually relevant information."
