Siloed, Efforts

Siloed AI Efforts Cost Companies 15% in Value Creation, Study Warns, as EU Regulation Looms

15.06.2026 - 00:13:23 | boerse-global.de

New KVP Institut study finds internal silos cause 70% of AI pilot projects to fail and reduce value creation by 15%, as local gains hurt system performance; EU regulations and training initiatives emerge.

Study: AI Silos Cost German Companies 15% Value, 70% of Pilots Fail
Siloed - Siloed AI Efforts Cost Companies 15% in Value Creation, Study Warns, as EU Regulation Looms 15.06.2026 - Bild: über boerse-global.de

A new study from the KVP Institut reveals that internal organizational silos are undermining artificial intelligence investments across German companies. The research finds that 70 percent of AI pilot projects fail to make a measurable contribution to overall corporate control, while isolated optimizations in individual departments lead to a value creation loss of roughly 15 percent.

The data points to a pattern of local wins at the expense of system-wide performance. While overall equipment effectiveness can improve by 22 percent within a single department, the same interventions cause variant throughput across the entire operation to fall by 12 percent. Similarly, sales forecasts locally gain 28 percent accuracy, but end-to-end delivery reliability drops by 8 percent. Leaders already spend 38 percent of their working hours on cross-departmental coordination, the study notes, yet the coordination gap remains a key drag on returns.

Alongside these internal challenges, external regulatory pressure is intensifying. On June 18, a series of webinars and expert presentations will roll out specifically for small and medium-sized enterprises and energy managers, focusing on the EU AI Act, the General Data Protection Regulation (GDPR), and the NIS-2 cybersecurity directive. Experts stress that security, liability, and organizational issues must be clarified early when deploying AI, including technical risks such as the OWASP Top 10 for large language models.

The regulatory push comes as AI capabilities continue to integrate into everyday business software. Microsoft unveiled MAI-Thinking-1 on Sunday, a model with over 35 billion parameters that the company says can reduce operational costs for AI applications by 90 percent. The technology is migrating into standard tools like Office and Windows 11, increasing the pressure on firms to build internal expertise for safe and effective use. A planned Windows security update on June 24 will refresh UEFI certificates on around 400 million computers, adding another compliance layer for IT departments.

Practical training offerings are emerging to address the skills gap across different functions. On July 2, a Berlin workshop will tackle HR use cases—recruiting, onboarding, and personnel development—using Microsoft Copilot and custom GPT adjustments. A webcast on June 16 covers AI-assisted account assignment in SAP systems, where pre-trained LLMs can automate booking suggestions and cut processing times by up to 80 percent. Training for automation platforms like Make.com and n8n is in rising demand, with workshops scheduled for July and October on multi-agent systems for managing complex business processes. The Fraunhofer Academy is additionally offering specialized formats for data-driven sectors such as energy and municipal heat planning.

en | boerse | 69541117 |