According to a survey from Red Hat, 87% of UK business IT decision-makers use agentic artificial intelligence (AI) systems, yet only a quarter of them have strong governance in place.
Currently, Red Hat claims that companies feel they do not have enough control over data, infrastructure and relationships with AI providers. Instead, their survey respondents seem to want governments to put in place stronger AI policies for industries.
Run by Censuswide, the survey questioned 500 IT decision-makers across the UK, France, Germany and Italy about AI sovereignty. The supplier concludes that there exists a control “gap” at organisations between their deployment of AI and their visibility into it. Whereas AI was, until recently, an aspiration, now most respondents view AI governance and regulation as a priority, now that AI deployment is more of a reality.
Most major AI suppliers are based in the US, making AI sovereignty a major concern for non-US governments. AI sovereignty means systems are run under the control of sovereign states, which allows for greater data privacy and control. As the technology advances, regulators are still playing catch-up, leaving some companies uncertain as to policy strategies.
Despite concerns over AI providers restricting access for customers to their deployed AI systems, only 67% of British IT decision-makers have an AI “exit strategy”. For many, this would have a “moderate to significant” impact on their businesses (43%).
“Executing a switch [from one AI technology supplier to another] without disruption remains difficult,” says Joanna Hodgson, UK country manager at Red Hat, which commissioned the survey. “To close that gap, enterprises need greater control over how and where AI runs, and a consistent way to govern fast‑moving technologies like agentic AI.”
Across the west European region, British respondents were most vocal in their demand for AI sovereignty regulation, with 89% of those surveyed wanting public policy and regulation to enforce the adoption of open source principles. Behind them, only 70% of French IT decision-makers and 72% of German respondents call for regulatory direction. This would mean stronger government guidelines on transparency, auditability and open source licensing.
For eight out of 10 decision-makers, this also involves letting companies have greater control over building these systems through open AI. While the demand for AI sovereignty increases, the gap between business and proprietary AI providers must be addressed, according to Red Hat.
In 2026, less than half of those surveyed (48%) have complete visibility of where their data is stored, processed and accessed. As such, letting businesses decide where AI is run could improve trust in the systems.
Hans Roth, senior vice-president and general manager for EMEA at Red Hat, pointed out that businesses want clear policy frameworks for AI. “Organisations are not looking for another closed, one‑size‑fits‑all stack,” he said. “They want the freedom to combine different models, accelerators and clouds while staying in control.”