Lünendonk® Study 2026: Agentic AI – from copilot to autopilot

For this study, Lünendonk & Hossenfelder surveyed 180 CIOs, IT directors, and line-of-business leaders across the DACH region. The results provide an unfiltered look at where companies actually stand today—and answer the questions currently top of mind for every IT leader: 

• Will AI agents replace ERP and CRM systems in the future, or will they simply orchestrate them? What decision-makers expect for their legacy environments.
• Why do so many agent projects get stuck in the pilot phase—and what exactly prevents them from moving into production?
• How much autonomy should an agent have? Learn how companies calibrate decision-making authority based on process risk.
• How AI agents turn legacy systems from bottlenecks into opportunities.
• What risks truly concern organizations—from shadow AI to vendor lock-in. 

In his accompanying expert article, Alfred Grünert (Global AI Lead at COSMO CONSULT) explains why custom process solutions no longer threaten IT budgets, thanks to new interface standards like the Model Context Protocol (MCP)—and why 60% to 80% of the effort in agentic AI projects is spent on building the data foundation. He shares three real-world consulting examples: from field sales in grocery retail and claims documentation in insurance to service technician support in mechanical and plant engineering.