AI has moved from experiment to enterprise priority, but it has brought a harder question: who controls the data that makes those systems work? New IDC research commissioned by Expereo finds 50% of UK enterprises want greater control over sensitive and strategic data, while 56% of UK technology leaders worry AI could create new security risks. For operational excellence leaders in Ireland, where GDPR and the EU AI Act apply, these are active risk management requirements today.

The IDC research is an operational resilience diagnostic as much as a technology survey. Its core finding is not that organisations are retreating from AI but that the foundations required to scale it safely need more attention. Data quality, governance, infrastructure readiness, and network visibility belong together in enterprise planning, not separately in legal and IT departments.

The scale of AI underperformance underlines the urgency. IDC found 62% of organisations describe their AI use as limited, and among organisations where AI underperformed, poor data quality was the leading factor at 51%, followed by costs exceeding ROI at 47% and AI underperformance at 46%. Organisations where AI exceeded expectations attributed success to strong skills, high-quality training data, and robust network performance: in other words, the operational foundations.

Digital sovereignty is the strategic frame through which organisations are responding. Some 29% of UK organisations view it as a top or high priority, and 27% rank it among leading drivers of technology investment. A further 26% say customer and partner demands to localise data are already influencing their approach, tying business performance to data control. Ben Elms, CEO of Expereo, stated digital sovereignty is driven by a desire for greater control over how data moves across cloud, network, and AI environments.

Network infrastructure is where the process improvement gap is most visible. IDC found only 5% of organisations say their networks are fully ready for future needs, yet 46% of successful AI implementations attributed results to network performance. Among organisations where AI failed, one quarter identified underperforming networks as a main factor. Elms noted that a sovereignty strategy without a network strategy is increasingly incomplete.

Three operational priorities emerge. First, map data flows before expanding AI: knowing how data moves across cloud, network, and AI environments is the foundation of governance and security. Second, treat data quality as an AI performance variable: 51% of underperformers cite it as the leading failure factor. Third, invest in network resilience alongside AI, not after it.

The IDC research confirms that AI value is not determined by the model alone but by the entire operational environment. Irish organisations that build data governance, network resilience, and security into their AI foundations will be best placed to scale with confidence and control.