A Netherlands-based deep tech startup developing real-time governance infrastructure that ensures autonomous AI systems remain governable, discernible, reliable, trustworthy, and accountable at scale.
Without proper governance infrastructure, Europe and the world face major safety, compliance, and ethical challenges as AI systems become increasingly autonomous. Organizations are deploying AI agents in high-stakes domains—healthcare, finance, infrastructure—without the ability to maintain meaningful oversight at machine speed.
Our purpose is to ensure the governability, discernability, reliability, trust, and accountability of autonomous AI at scale. We believe that as AI systems gain autonomy, humans must retain the ability to influence the decisions they delegate—not through post-audit checks or slow development cycles, but in real-time, as reasoning unfolds.
Our mission is to enable enterprises to delegate critical work to autonomous systems without losing control, traceability, or compliance—while keeping up with the rapid pace of AI adoption.
We make AI governance intrinsic—embedded within cognition itself as it happens. Not waiting for post-audit checks. Not tweaking rules at development cycles. Governance at runtime, auditable end-to-end, with humans retaining meaningful control over autonomous decisions.
Traditional AI governance is reactive—auditing decisions after they've been made. We provide real-time cognitive oversight, enabling intervention before consequential actions are executed.
Governance isn't bolted on—it's embedded within AI reasoning itself. Our Steward Agents observe decision-making as it unfolds, providing continuous oversight without operational latency.
Humans maintain meaningful control through intelligent alert systems and intervention mechanisms. The system amplifies human judgment rather than replacing it.
Built on the Reflection Protocol (RDP), an open standard for AI governance. We believe trustworthy AI requires transparency—open specifications that anyone can verify, audit, and improve.
Domain experts, regulators, and stakeholders can directly participate in AI oversight through machine-readable Governance Blueprints—portable, versionable policy frameworks.
Our cognitive governance architecture combines patent-pending technology with participatory principles, creating infrastructure that's both technically innovative and community-driven.
Every decision is explainable and auditable. We build on open protocols rather than proprietary black boxes, ensuring visibility into how governance actually works.
AI safety should be a collective effort, not locked in corporate silos. We're committed to open research, community collaboration, and sharing governance knowledge as public infrastructure.
Technology should augment human judgment, not replace it. Our systems ensure humans remain in meaningful control, with the ability to intervene when it matters most.
Compliance isn't enough—we build for real safety. Our governance mechanisms create actual oversight, not compliance theater, with runtime intervention capability.
We're committed to advancing AI safety as a field, not just as a business. That's why we're collaborating with AI Safety Camp to empirically test the Cognitive Governance Protocol (CGP) and open-source fundamental aspects of our framework.
Our research explores whether governance can function as portable, participatory infrastructure—creating reusable building blocks for AI safety that the entire community can benefit from. All research findings, datasets, and governance templates will be released openly.
We believe that by treating governance as public infrastructure—like internet protocols for communication or security—AI safety can scale with autonomy. Our open research commitment ensures that governance knowledge becomes portable across domains instead of siloed within individual organizations.
Based in the Netherlands and working globally to build the governance infrastructure for autonomous AI. Whether you're an enterprise exploring AI governance, a researcher interested in our open protocol, or a potential partner, we'd love to hear from you.