IBM Report: Most EMEA Executives Don’t Fully Understand Their AI Dependencies
Summary
An IBM Institute for Business Value report, in partnership with Oxford Economics and published on June 17, 2026, reveals that 90% of executives in EMEA lack a full understanding of their organizations' AI dependencies across vendors, models, and infrastructure. Globally, only 9% of 1,000 surveyed executives believe they have excellent visibility, and 71% find it difficult to switch their primary AI provider or model. This lack of insight raises significant concerns regarding unpredictable costs, operational outages, and vendor lock-in, especially as AI's influence on operational decisions is projected to rise from a quarter to nearly half by 2030. The report redefines AI sovereignty as maintaining control amidst changing conditions, advocating for "selective AI sovereignty" to prioritize governance for high-impact applications like fraud detection and risk management, rather than attempting full stack ownership. Organizations with stronger AI control frameworks demonstrate greater resilience.
Key takeaway
For Directors of AI/ML managing expanding AI deployments, you must gain comprehensive visibility into your organization's AI dependencies across vendors, models, and infrastructure. Implement "selective AI sovereignty" by prioritizing robust governance for critical systems like fraud detection and risk management. This approach protects your operating profit and ensures resilience against potential vendor changes, technical shifts, or regulatory pressures, preventing costly outages and vendor lock-in.
Key insights
Many EMEA executives lack critical visibility into AI dependencies, risking costs and operational stability.
Principles
- AI sovereignty means maintaining control, not just ownership.
- Fragmented AI governance increases operational and financial risk.
- Selective control enhances resilience for critical AI systems.
Method
The report introduces "selective AI sovereignty," focusing governance efforts on high-impact applications like fraud detection and risk management, while allowing flexibility for lower-risk areas.
In practice
- Prioritize governance for high-impact AI applications.
- Map AI vendor, model, and infrastructure dependencies.
- Assess potential costs of primary AI provider outages.
Topics
- AI Dependencies
- AI Sovereignty
- AI Governance
- Vendor Lock-in
- Operational Resilience
- EMEA Business
Best for: CTO, VP of Engineering/Data, Executive, Director of AI/ML, Consultant
Related on AIssential
Editorial summary, takeaway, and curation by AIssential. Original article published by TechRepublic.