Profile
Founder
Malcolm is building GraphCentric around the idea that application data should live in a shared, governed semantic layer rather than inside isolated application silos. His work focuses on RDF, web-native resource design, access-controlled updates, and making enterprise data safely usable by both people and AI agents.
Writing
AI projects need more than larger models: they need reliable business context. This article explains why context graphs and semantic layers help humans and AI agents access connected, traceable information for better decisions.
Read article
Link relations give machines explicit cues for how to move through a website without relying on brittle, hardcoded APIs. This article explains why the original web model was already agent-friendly and why that matters again in the age of AI agents.
Read article
OAuth2 scopes provide a practical way to limit what AI agents can do on behalf of users in production systems. This article shows how a mature web standard can become an enforceable safety boundary before an agent gets anywhere near destructive operations.
Read article