A survey commissioned by software company SAS indicates that while agentic or "agent-based" AI systems are increasingly used in business decision-making, executives' trust has not kept pace with the technology's growing autonomy. The study polled 2,700 decision-makers across 28 countries; the findings were reported by J.D. Capelouto.
Key findings
- Nearly 90% of respondents said agentic systems play some decision-making role within their organizations.
- Only 66% said they trust agentic AI.
- By comparison, roughly three-quarters (about 75%) said they trust generative AI more broadly.
Taken together, these figures suggest that businesses are implementing AI tools faster than leaders are developing confidence in those systems.
Why are people overriding AI decisions?
Respondents identified insufficient explanation of how the AI reached its conclusion as the top reason for overriding an AI decision. This explanation gap was cited more than twice as often as the output simply being wrong.
This indicates that lack of transparency and interpretability is a larger barrier to adoption than accuracy errors alone: even correct outputs may be questioned if users do not understand the reasoning behind them.
Why this matters
Organizations are rapidly integrating AI into operations, but effective adoption and scaling depend on users understanding and trusting how those tools work. The survey suggests that improving explainability and educating users should be priorities for business and technology leaders if AI systems are to be relied upon for routine decision-making.
Conclusion
The SAS-commissioned survey shows a gap between the pace at which agentic AI is adopted and the level of executive trust in those systems. Addressing explainability and user confidence will be important steps to ensure AI deployments deliver sustained value in business contexts.
(Reported by J.D. Capelouto; survey commissioned by SAS.)
