Companies are adopting AI at a rapid speed. AI tools and agents now drive business outcomes at an ever-increasing scale, with an ever-increasing autonomy. Keeping track of the accompanying technical and regulatory complexity is becoming a key issue for AI-enabled organizations. Our portfolio company LatticeFlow AI is closing exactly that gap: their newly launched AI Platform is purpose-built for the agentic era, connecting governance frameworks directly to technical controls. It enables enterprises to understand, control, and govern AI risk with evidence, continuously. Congratulations to Dr. Petar Tsankov and the entire LatticeFlow AI team for this milestone.
AI governance has a performance problem. Policies written. Frameworks adopted. Audits passed. Boxes checked. And somewhere in the background, an AI system is doing something no one fully understands, because the governance layer was designed to satisfy a reviewer, not to control a risk. That's governance theatre. It looks rigorous. It's not defensible. The shift happening right now in AI teams: moving from documentation-based governance to technical evidence. Not "we have a policy against this." But: "here's the test, here's what the score actually means for our enterprise risk, here's what changed last week." AI governance is evolving beyond policies and processes. It’s becoming a technical control function for managing AI risk. As Dr. Holger Harms, Head of Banking Innovation Lab at Swisscom, put it: "Agentic AI is fundamentally changing what effective governance requires. As AI systems gain the ability to reason, use tools, and take autonomous actions, policies and periodic reviews are no longer enough. Organizations need to continuously discover where AI is being used, evaluate how it behaves, and govern it through technical risk controls that evolve alongside these systems. This is especially critical in banking, where innovation must scale alongside trust, resilience, and regulatory accountability. LatticeFlow AI is making important contributions to the evidence-based control of agentic systems.”