Plenty of ideas, but no clear priorities
Use cases and experiments emerge more quickly than sound decisions about which projects are economically viable and technically feasible.
Individual use rather than process impact
AI speeds up individual tasks, whilst workflows, handover processes and systems remain largely unchanged.
Successful prototypes do not make it into production
The AI function works in principle, but architecture, integration, quality and operations have not yet been adequately taken into account.
Usage and costs are becoming difficult to manage
Models, providers and technical access points are evolving in parallel, making it difficult to achieve transparency regarding usage and interdependencies.
AI remains separate from existing products and systems
New functions are tested in isolation, rather than being meaningfully integrated with software, data and real-world user processes.