Blog
Blog
178 posts in Blog.
From Harness to Graph: What the Next Wave of AI Engineering Means for Industrial Operations
In July 2026 the AI engineering conversation moved again — from the harness to graph engineering: designing an AI system's topology as an explicit, inspectable graph of agents, tools, deterministic logic, and governance gates. It turns out industrial operations has been building exactly this for over a decade.
NGNic Gould
What 50 Analyst Research Appearances Reveal About XMPro
From digital twins to Agentic Operations, the pattern behind XMPro's 50 analyst research appearances since January 2024 tells a bigger story than the number.
Wouter Beneke
An Ontology Is Not a Graph Database
Why industrial context needs a system of fit-for-purpose data stores, standardized and governed, before an agent can act on it safely.
Pieter van Schalkwyk
Who Owns Your Context Layer?
France's intelligence agency dropped a platform it could not itself turn off. Any operator building an industrial context layer faces the same exposure, and the consolidation into OEM stacks makes it urgent. The procurement question has changed from which platform has the best features to: whatever happens to this vendor, what do I still own?
Pieter van Schalkwyk
288 Years of Research Behind MAGS
The oldest idea in the MAGS codebase was published in 1738. A tour of the 288-year research lineage built into MAGS — every architectural layer traced to the published work it implements, cited at the line of code.
Pieter van Schalkwyk
Why Industrial AI Agents Don't Need to "Reason" Like ChatGPT
Industrial AI agents need more than a reasoning model — they need a governed decision architecture. Pieter van Schalkwyk explores XMPro's OODA-inspired Cognitive Decision Loop (ORPA: Observe, Reflect, Plan, Act) that lets MAGS work like virtual operators within bounded authority, leaving an auditable Decision Trace.
Pieter van Schalkwyk
Your AI Agent Needs a Harness — Here's What That Means for Industrial Operations
The AI industry has learned a hard lesson over the last eighteen months: the model isn't the hard part — the runtime environment around it is. That infrastructure now has a name: the agent harness. Here's what it means for industrial operations, where wrong answers have physical consequences.
NGNic Gould
XMPro CEO Pieter van Schalkwyk Speaks at CIIT Mining Conference in Peru on Agentic Operations and Building Trust in Industrial AI
Pieter van Schalkwyk presents on agentic operations and building trust in industrial AI at the CIIT Latam Congress in Lima, Peru.
Pieter van Schalkwyk
Beyond Predictive Maintenance: How Causal Analytics Eliminates Recurring Failures
Beyond Predictive Maintenance: How Causal Analytics Eliminates Recurring Failures Author: Marcos Augusto Burgos Saavedra The Early Detection Approach A critical pump in the flotation process at a...
MAMarcos Augusto Burgos Saavedra
Beyond OpenClaw: Why Industrial Agents Need Bounded Actuation
Pieter Van Schalkwyk CEO at XMPRO This article originally appeared on XMPro CEO's Linkedin Blog, The Digital Engineer I've worked on autonomous operations for years, back when agents weren't sexy or...
Pieter van Schalkwyk
Sunday Morning Strategy: How I Use AI for Continuous Tactical Calibration
Pieter Van Schalkwyk CEO at XMPRO This article originally appeared on XMPro CEO's Linkedin Blog, The Digital Engineer Sunday morning, February 15, 2026. I'm reading The Australian over coffee when an...
Pieter van Schalkwyk
The 'AI Wants to Kill You' Story Shows Why Architecture Matters More Than Alignment
Pieter Van Schalkwyk CEO at XMPRO This article originally appeared on XMPro CEO's Linkedin Blog, The Digital Engineer Mark Vos 's adversarial testing of an AI assistant made headlines across...
Pieter van Schalkwyk
The Decision Traces You Have and the Ones You Need
Pieter Van Schalkwyk CEO at XMPRO This article originally appeared on XMPro CEO's Linkedin Blog, The Digital Engineer Jaya Gupta argues that decision traces are the moat in her recent follow-up...
Pieter van Schalkwyk