Europe’s debate about technology happens on two tracks that are actually one. Talk about infrastructure, dependency, and competitiveness on one side; rights, privacy, and dignity on the other. Whoever controls the technology decides what it does to people.
The continent routinely gets labeled a laggard in artificial intelligence. That framing misses the point. Speed measures nothing about direction. The harder questions involve which capabilities Europe wants for itself, built on what assumptions, and what standards AI must meet before entering public life.
Every AI procurement answers those questions. No system arrives without values. The builder’s values get baked in, and regulation alone rarely secures them. Lasting commitments require structure, not just good intentions, because promises survive when companies bind themselves to them.
ChapsVision, which develops data-intelligence and AI software for governments and enterprises, pursues what its founder calls a double mandate. Olivier Dellenbach launched the company alongside the HappyCap Foundation, which supports families raising children with cognitive disabilities. He frames both as one long-term bet on autonomy: for those families, and for the institutions his firm serves.
The setup matters because HappyCap is designed to eventually become ChapsVision’s main shareholder. That arrangement would fund the foundation’s mission permanently while shielding the company from hostile takeovers. Dellenbach says he did not build the company to flip it once traction arrived.
His philosophy traces back to parenting a child with a disability. Responsibility there means years of adaptation and accompaniment, not quick fixes. The goal remains autonomy, calibrated to each person’s situation. Dellenbach sees the same dynamic with customers. His teams aim to make organizations more capable with their own data, not permanently dependent on outside technology.
For public institutions, procurement choices carry consequences that outlast any software license. Architecture, interoperability, data ownership, and security determine whether an agency grows more skilled or more captive. As AI evolves rapidly, governments should avoid systems that foreclose future options. Vendor lock-in, Dellenbach argues, equals lost freedom.
ChapsVision maintains an independent Ethics Committee to scrutinize deployments involving human rights, democratic principles, or geopolitical risk. The committee does not block innovation. It asks whether a project should proceed, and under what conditions.
The company’s competitive thesis: European values only become advantages when embedded in technology that performs at a global level. Transparency, accountability, privacy, and strategic autonomy mean little if the products cannot match alternatives on quality. Europe need not recreate every layer of the US and Chinese stacks. The opportunity lies in application software: turning models and data into operational intelligence for organizations.
In defence and crisis management, agentic AI should expand operational capacity without surrendering decision authority. Agents can gather information, connect signals, analyze volumes, and prepare scenarios. But consequential actions demand stronger governance. Clear boundaries must define what agents do autonomously, what needs human validation, and how every action gets traced.
The principle remains straightforward: human in the loop, always. Not to slow AI down. To make it trustworthy enough for the highest-stakes environments.
Europe possesses the talent, scientific base, and industrial capacity to build this layer. The challenge lies in converting those assets into lasting companies that compete globally while keeping critical institutions under sovereign control. That is where ChapsVision aims to operate: European in values, global in ambition, and trusted where failure carries the steepest costs.
