IDC: EMEA CIOs Need to Rebuild Data and ROI Models to Move AI to Production

IDC believes that corporate AI implementations in EMEA have stalled not because of a cooling toward the technology, but due to poor preparation for scaling. To move projects out of pilot stage, CIOs need to simultaneously reconsider the economics, data architecture, and how new tools are integrated into teams' daily workflows.

Why Pilots Aren't Growing

Over the past year and a half, companies in Europe, the Middle East, and Africa have actively invested in large language models and classical machine learning, betting on significant operational benefits. But now boards of directors are beginning to slow down some programs, reduce scope, or shift investment focus. According to IDC, it's not a loss of interest in AI, but rather that business requires stricter financial justification against competing IT priorities and macroeconomic pressure.

The numbers are stark: only 9% of organizations in the region have managed to get measurable business results from most of their AI projects over the past two years. The remaining 91% have not necessarily encountered technical failure, but are stuck in the gray zone between demo and real production. Projects don't die loudly—they lose momentum, budget, and internal support without becoming a sustainable corporate function.

Where ROI Breaks

One of the main mistakes is evaluating AI using old procurement logic, where the value of new software is measured through headcount reduction or direct license savings. For generative models, this often doesn't work. Their effect manifests indirectly: through reduced risk, accelerated specialist work, decreased downtime, and new revenue sources. If a company looks only at obvious costs, a promising pilot almost inevitably loses the budget battle. CIOs essentially need to redescribe the economics of implementation and audit not just the model, but the entire chain of its operation:

• prevented losses, such as from downtime or errors
• improved team productivity and shortened cycle times
• new revenue scenarios and digital services
• full cost of inference, storage, integrations, and support
• security, compliance, and data control expenses

The next barrier emerges when transitioning from a cloud sandbox to a corporate environment. A pilot can quickly launch on APIs and test data, but production requires constant compute, live pipelines, monitoring, and integration with legacy systems. When modern vector databases need to connect to old Oracle or SAP systems, architectural gaps surface immediately. For RAG scenarios, you need clean, annotated, and properly classified data; otherwise, answer quality drops and hallucinations increase. Meanwhile, inference costs and model fine-tuning bills grow, which already need to be explained line-by-line to finance teams.

Architecture and People

IDC specifically emphasizes that requirements for data protection, cybersecurity, and model explainability in Europe don't necessarily hinder scaling. On the contrary, companies that establish governance and control rules from day one move faster. Protection against prompt injection, clear access boundaries, documentation of model decisions, and data control increase the base project cost, but simultaneously make the system suitable for real corporate use and strengthen customer trust.

Equally significant resistance emerges at the employee level. An AI solution can be technically functional and still fail to take root if it breaks familiar processes or demands too abrupt a shift from teams. Therefore, CIOs need to invest not just in models, but in retraining, change management, and implementation design around how people actually work.

Automated contract review makes sense when lawyers spend less time on routine work and more on negotiation and complex risks. Against this backdrop, the role of CIO is changing: according to IDC data, 42% of top executives in EMEA expect the CIO to lead digital and AI transformation with a focus on creating new revenue sources.

What It Means

For EMEA, the phase of "let's just try AI" is ending. The next stage will be won not by companies with more pilots, but by those who can prove their value, prepare data, and embed models into real business processes without a gap between technology, finance, and people.