About UNIT2
Built for operations.
Not for the AI hype cycle.
UNIT2 is a Dutch AI company building private, model-agnostic AI workforces for organisations where knowledge continuity, productivity, data and accountability genuinely matter.
Our conviction
The next organisation has two execution layers: human and AI-first.
European organisations need to increase output while experienced people become scarce and expensive. A chatbot beside every desk does not change that economic reality; a demonstrably governed process layer can.
UNIT2 therefore builds a second organisation beside the first: roles, work queues, company knowledge, authority, controls and decision history. The two layers initially work in tandem. As the digital layer proves itself, it can take over suitable execution during growth, illness, retirement or natural attrition. People retain direction, relationships and exceptions.
Process before technology
Implementation begins with tasks, exceptions, quality and ownership—not a model demo.
Private and model-independent
We design infrastructure, agents, model choice and processes as one replaceable, coherent system.
Control as a building block
Governance, data access, audit and human control shape the architecture from day one.
The team
Small, senior and directly involved in delivery.
UNIT2 deliberately works with a compact core. Process design, systems architecture, European governance and commercial viability therefore stay close to the same work floor.

Ronald J. Kraft
Combines systems architecture with years of experience in engineering, offshore robotics and autonomous operations. Focuses UNIT2 on private technology that demonstrably works in real business processes.

Sailor K. Kraft
Connects computer science and neuropsychology in secure data architecture, agent behaviour and governance. Translates European rules into practical technical controls.

Jeroen Dam
Translates complex AI architecture into concrete process layers, business value and executable implementation programmes.
Work with us
Let us define your first AI-first process layer.
The best first conversation starts with real information about volume, knowledge, systems, exceptions, data boundaries and desired human control.