PROBABLY RIGHT IS THE WRONG TARGET
Knowledge Graphs
In regulated work, an approximate answer is a liability. The target is not a more capable model but deterministic integrity: the same answer every time, with every claim traceable to its source.
Insights
THE AXIS HAS ROTATED
AI Strategy
Build versus buy is no longer the real question. When the model is a commodity input on both sides, what matters is whether you own a substrate worth running a model against — and whether it compounds faster than a rival can copy it.
THE USER INTERFACE IS NOW A COMMODITY
AI Strategy
The legal AI workspace is being compressed from four directions at once. When the surface stops being the product, the defensible layer underneath is the ontology.
NOT ALL TOKENS ARE CREATED EQUAL
Engineering
Most RAG pipelines quietly assume every token carries equal expected information value. That frequentist assumption compounds at every layer — and it shows up on the bill.
THE DATABASE I CHOSE WHEN EVERYONE ELSE CHOSE POSTGRES
Engineering
Choosing infrastructure under regulatory constraint: why a graph-native, multi-model engine beat the default relational choice for adversarial financial disclosure — and what that decision actually costs.
I AM THE ENTERPRISE
Perspectives
Marx built his theory on the separation of ownership from labour. AI-native tooling did not hand the factory to the workers — it dissolved the separation. What emerges is embedded technical capital.
WHY KNOWLEDGE GRAPHS ARE THE BACKBONE OF ENTERPRISE AI
Knowledge Graphs
In a world drowning in unstructured data, knowledge graphs provide the connective tissue that transforms raw information into actionable intelligence. Here's why every serious AI strategy starts with a graph.
BUILDING AN AI STRATEGY THAT ACTUALLY WORKS
AI Strategy
Most AI strategies fail not because of technology, but because of architecture. A conversation about the frameworks that separate transformative AI programmes from expensive experiments.
THE ARCHITECT'S VIEW: MODERN DATA ARCHITECTURE
Engineering
From data mesh to knowledge fabric — the architectural patterns that enable organisations to build AI systems that scale, adapt, and deliver measurable value.
WHAT ROAD CYCLING TAUGHT ME ABOUT SYSTEMS THINKING
Perspectives
The peloton is a complex adaptive system. Every ride is an exercise in efficiency, interdependence, and marginal gains. The parallels to enterprise architecture are striking.
WHERE NEURAL NETWORKS MEET KNOWLEDGE GRAPHS
Knowledge Graphs
The convergence of statistical AI and symbolic reasoning is reshaping what's possible. Neuro-symbolic architectures are the next frontier — and knowledge graphs are at the centre.
THE AI TRANSFORMATION BLUEPRINT
AI Strategy
Transformation isn't a technology project — it's an organisational one. A practitioner's guide to the human, structural, and technical dimensions of real AI transformation.
ONTOLOGY DESIGN: THE CRAFT BEHIND THE GRAPH
Knowledge Graphs
A well-designed ontology is the difference between a knowledge graph that works and one that doesn't. The principles, patterns, and hard-won lessons from the field.
GRAPH-POWERED RAG: BEYOND VECTOR SEARCH
Engineering
Retrieval-Augmented Generation is powerful, but vector search alone misses context. Graph-powered RAG adds structured reasoning to make LLMs genuinely useful in the enterprise.