A major Dutch ministry was experiencing structural performance issues within its existing AI systems. Constraints in the underlying infrastructure, combined with fragmented document and knowledge flows, made it difficult to scale AI applications reliably and deploy them to support further automation. At the same time, pressure was increasing to demonstrate compliance with national and European AI legislation governing the responsible and transparent use of AI.
The organisation aimed to improve the efficiency of its AI systems, strengthen governance and operate in demonstrable compliance. The central question was therefore twofold. How can AI performance be enhanced, and how can those improvements be embedded sustainably and responsibly within the organisation?