Deploying AI agents in production? This series is about the engineering of AI governance: the technical controls, runtime guardrails, and audit infrastructure that turn policy into something you can actually ship. Sandboxing, policy-as-code, agent-native telemetry, least-privilege identity, all with real code examples using Microsoft’s open-source Agent Governance Toolkit, honest gap analysis, and no vendor spin.

The EU AI Act is the regulatory backbone: its Article 50 transparency rules become enforceable on August 2, 2026, while the Digital Omnibus agreement moves the high-risk obligations to December 2, 2027 (stand-alone systems) and August 2, 2028 (product-embedded systems), subject to final adoption. But the through-line here is broader: governance you can deploy, in Europe and beyond. Policies are not controls. This series is about the controls.

Written by Carlos Hernandez, founder of GenAI Gurus, Europe’s GenAI practitioner community, a contributor to the Microsoft Agent Governance Toolkit, and a contributor to AgenTrust, a new open-source stack for governed, provable agent execution, where I focus on embodied AI in industrial environments. For a curated index of EU AI Act tools, official sources, and templates, see Awesome EU AI Act.

The Series

# Post Status
1 EU AI Act for AI Agent Developers: A Practical Compliance Checklist Published
2 How to Run Coding Agents Safely in the Enterprise Published
3 When Autonomous AI Agents Control Robots: The Real Threats No Safety System Detects Published
4 Verifiable Trust for AI Agents That Control Robots: A Working Example with AgenTrust Published
5 The Cloud Can Prove It. The Robot Can’t. Published
6 Proof of Outcome: Did the Robot Actually Do It? Published
7 Trust at Fleet Scale: The Skill That Spreads Published
8 Trusting the Skills You Didn’t Write Published
9 Signed Is Not Trustworthy: Grading the Evidence Behind a Robot’s Action Published
10 Code Is No Longer the Bottleneck. Verification Is: A Survey of AI-Native Software Delivery in 2026 Coming soon
11 From NIM to Jetson: A NeMo Guardrails Configuration Pack for Production Inference Coming soon
12 Open Weights, Real Obligations: Governing GPAI Models You Deploy but Didn’t Train Coming soon
13 Sovereign AI Infrastructure: Governance Patterns for On-Prem and European Cloud Coming soon
14 The Contributor Journey: Building an Open-Source Agent Governance Layer Coming soon

Posts

subscribe via RSS