EN 18286 Is Here: Presumption of Conformity Is Not Runtime Conformance EN 18286:2026 gives providers of high-risk AI systems a recognized route to a presumption of conformity with the Article 17 quality-management requirements it covers. It does not decide whether an agent’s next tool call is authorized. Presumption of conformity is a property of your quality management system. Runtime conformance is a property of your actions. You can document the first. You must enforce the second, before an action executes.
The Controls Have to Sit Outside the Agent An agent can be told what it may not do, but only a control outside the agent can stop it. On October 10, 2026, Satya Nadella, CEO of Microsoft, argued that traditional software let us trace behavior to a specific code path, while that kind of mechanistic understanding does not carry over to agentic systems. Those agents are already being connected to sensitive data and given the ability to take mission-critical actions. He wrote: “We simply can’t outsource responsibility for what intelligence does on our behalf. A model provider’s assurances do not relieve us of that responsibility.” His central instruction was to “separate the supply of intelligence from the authority over it.” This post takes that framing and follows it where it leads for organizations deploying agents in production: the controls, evidence, and ability to stop an agent must be independent of the intelligence being governed.
When Agents Can Move Money, the Prompt Is Not a Control We keep writing instructions for our agents as though instructions were enforcement. We tell an agent what it may not touch, and we describe the result as a control. We phrase a boundary in a prompt, and we treat that phrasing as authority. But a prompt is not a control, and a prompt is not an authority. What we write in a prompt shapes what the agent intends to do; it does not govern what the agent is able to do. Instruction does not control actions. Those three statements are the entire argument, and payments is where getting it wrong costs the most. In a recent J.P. Morgan Payments article, Zack Anderson, Chief Data & Analytics Officer for Payments and Global Banking, states the principle plainly: a prompt expresses intent; it does not enforce authority. That framing is correct, and it carries a consequence that most agent programs have not yet absorbed. If the prompt is not the control, then something else must be, and that something has to exist as infrastructure in every agent project rather than as a review step bolted on later. This post is about why that layer has to exist at all, and the best way to see the problem is to work through an example from the article itself.
NVIDIA OpenShell Is a Major Step for Agent Security. Runtime Governance Is the Next One. NVIDIA OpenShell addresses a problem that every serious AI deployment now faces: an agent needs access to systems and data, but it cannot be trusted to govern that access itself. OpenShell provides an open-source runtime boundary around the agent. It controls what the agent can reach, what it can access inside its environment, and which policy changes can take effect. That is foundational infrastructure. But runtime isolation is not the same as business authorization. An enterprise also needs to know what the agent is capable of doing, what each operation means, and whether a specific action is authorized at the moment it is attempted. That requires three additional layers: management, design-time action-surface mapping, and deterministic runtime authorization.
AI Governance Isn’t a Chore Problem. It’s an Architecture Problem. Four in five senior business decision-makers are spending more of their day managing AI risk, and their working hours are up 26% on average, according to the September 2026 OneTrust survey cited by CIO.com. That is not evidence that enterprises need more governance meetings… it is evidence that governance is being performed in the wrong place. Leaders are not governing. They are triaging incidents, chasing unapproved tools, reconstructing decisions, and reviewing actions after they occur. The chore is real, but it is a symptom. The condition is architectural. Enterprises built governance as a human-review activity bolted onto deployment. They then aimed that process at a growing population of non-deterministic systems operating at machine speed. No amount of additional analyst capacity fixes that design. The answer is to move governance to the decision point before execution, enforce policy deterministically, and reserve human attention for exceptions. This is the distinction: Not after-the-fact review, before-execution authorization. Not more alerts, fewer uncontrolled decisions. Not a reviewer for every action, a person for every meaningful exception.