Knowledge
Understand, design, secure, situate
This is where the content of agent-007.ai comes together. Four movements — understand, design, secure, situate — lead through the same knowledge base, depending on your question and prior knowledge.
What is agentic AI?
A Chatbot answers. A Copilot suggests. An Agent plans, picks tools and acts — and it is precisely this step from “answering” to “acting” that shifts the risk from the wrong answer to the unwanted action. That is why everything on this page revolves around the architecture that makes acting reliable, auditable and bounded.
Paths through the knowledge base
The seven missions
Seven fields where AI readiness is decided — each with a guiding question, patterns, anti-patterns and primary sources.
Open ArchitectureEnterprise AI Control Plane
The reference architecture: five layers and one cross-cutting Control Plane.
Open OrientationArchitecture radar
What to establish now, to trial, to scrutinise or to deliberately avoid — with reasoning.
Open InteractiveAI-Readiness-Pulse
An honest self-assessment across the seven missions — evaluated entirely in the browser.
Open ReferenceGlossary
Core terms from Agent through RAG and MCP to Control Plane, explained concisely.
Open TrustSources & methodology
How the content is created, checked and maintained — and where its limits lie.
OpenAsk instead of browse
If you would rather ask than read: Nova-7 answers questions drawn from exactly this content and points you to the relevant pages and sources.