New global research from Denodo has revealed a widening “trust gap” that could slow the adoption of agentic AI, as organisations struggle with fragmented data environments, governance challenges, and lack of real-time access.
The findings, published in the AI Trust Gap Report based on a survey of 850 global executives, highlight a growing disconnect between enterprise AI ambitions and the data infrastructure required to support them. As AI systems evolve from chatbots to autonomous agents capable of making decisions and triggering business processes, the reliability and timeliness of data have become critical.
According to the report, 66% of organisations consider real-time or near real-time data access essential for AI systems to be trusted. However, achieving this remains a challenge, with 67% of respondents citing difficulties around data security and access controls both crucial for ensuring safe and compliant AI operations.
“AI is rapidly shifting from systems that merely answer questions to systems that take autonomous action, and this transition changes the data requirement entirely.” Richard Jones, Vice President and General Manager for Asia Pacific and Japan at Denodo.
Data complexity is another major hurdle. Around 42% of organisations reported sourcing data from more than 400 different systems for their AI initiatives, significantly increasing the risk of inconsistency and latency. Additionally, nearly 60% of respondents said they face performance bottlenecks when handling the intensive workloads required for large-scale AI deployments.
The study also found that 63% of organisations struggle to identify and prepare the most relevant and trustworthy data for AI consumption, underlining the importance of data quality and contextual accuracy in driving meaningful AI outcomes.
Experts warn that as AI systems move from providing insights to taking action, the margin for error narrows considerably. In agentic environments, inaccurate or outdated data could directly impact business decisions and operations, raising the stakes for data governance and infrastructure readiness.
The report concludes that the trust gap is not rooted in AI models themselves, but in the underlying data architecture. To scale AI effectively, organisations must modernise their data ecosystems, ensuring seamless integration, real-time availability, and robust governance frameworks.
As enterprises accelerate their AI strategies, bridging this data trust gap will be essential to unlocking the full potential of autonomous, decision-making systems.