One of the simplest yet most powerful ways to structure information is with a triple.

A triple is exactly what it sounds like: a unit of data broken into three parts - subject, predicate, and object. Think of it like a bite-sized fact. For example:

🔹 Subject: Big Ben
🔹 Predicate: is located in
🔹 Object: London

This seemingly basic structure is the foundation of knowledge graphs - those sprawling networks of interconnected facts that power everything from Google Search to a new breed of AI assistants.

In the AI arms race, data isn’t just fuel - it’s the architecture for the intelligence you train. Yet most enterprises still rely on 20th-century data architectures for 21st-century intelligence. Your CRM is a vault of customer interactions, your ERP tracks orders, and your analytics tool crunches numbers - each a walled garden. AI is meant to be the brain that connects them all, but it can’t - because these systems weren’t designed for AI.

In a simple graph, an edge between two nodes just means "these things are connected." In a knowledge graph, the edges say how and why they are connected.

Let’s expand our example. Suppose Alice isn’t just a person - she’s a doctor. She works at a hospital. That hospital is located in London and specialises in cardiology. Instead of an undifferentiated mess of connections, we now have semantics - explicit labels that tell us what each node and edge means.

This is what turns a graph into a knowledge graph: it captures relationships, categories, and meanings. It understands that a person isn’t the same as a company, and that "works at" is different from "has visited."

What does DeepSeek mean for enterprises? The release of DeepSeek’s R1 model has caused shockwaves. This feels like a pivotal moment - demonstrating how constraints can drive innovation while also hinting at the economic and geopolitical 'mega-event' looming on the horizon. Beyond its clever efficiency gains, R1 underscores two critical trends that enterprises should heed: the rise of open-source AI and the irreplaceable value of high-quality data.

For the last couple of years, many dismissed AI as another overhyped tech trend. But that phase is coming to an end. Organisations are racing through the denial stage, accepting that AI is real, and now grappling with the sheer scale of its impact. And for knowledge-intensive industries - as well as those providing software as a service - that realisation often leads to full-blown panic.

As AI accelerates through the economy, organisations with poorly integrated data systems will begin to show cracks. Disparate but entangled data quality issues will lead to unreliable AI insights and a loss of trust. Within a ten-year timeframe, many organisations may crumble under the strain of their fragmented infrastructures, losing relevance as their specific intelligence fades into the background intelligence of larger foundational models.

OpenAI’s o3 model has set new highs in significant benchmarks—and that's a game-changer for all of us. If AI can reason, code, and excel in maths and science, it’s only a matter of time before it starts reshaping tasks critical to nearly every business. Let’s dive into how o3 performed on key benchmarks:

Here are The Knowledge Graph Guy’s predictions for Knowledge Graphs in 2025!. These trends underscore the increasing strategic importance of Knowledge Graphs as a cornerstone for GenAI adoption in 2025.

An ontology is like a map of your organisation’s knowledge, built on the concepts and relationships that define your domain. Think of it as your company’s DNA—a compressed representation of what you know and how you think. When developed properly, your ontology doesn’t just support your AI; it becomes a core piece of your intellectual property.

Models aim to capture the essence of a system, preserving its logic and structure in a way that enables reasoning, prediction, and deeper understanding. A simulation, on the other hand, mimics the observable behaviour of a system without necessarily representing its underlying structure or logic. While both are valuable, they serve different purposes: models aim to explain and understand, whereas simulations focus on reproducing and predicting.

A Virtual Knowledge Graph effectively becomes the Turing Machine’s “tape”—a dynamic, interconnected memory that the LLM can read from and write back to. By linking facts with URLs and using ontology for generalisation and compression, this graph provides an ideal read/write memory structure for the LLM. Combined in a neural-symbolic loop, this LLM+Graph system could even be Turing complete.

For Artificial Intelligence to be successfully integrated into an organisation, it must build upon and amplify the existing unartificial intelligence within that organisation.

Artificial Intelligence works through data and computation. To use AI to improve Organisational Intelligence, it must make the organisation's internal model of the world more cohesive. It must achieve this by enhancing collaboration and knowledge sharing between people through data and computational insights. It must empower Collective Intelligence to achieve Organisational Intelligence.

Metadata is essentially "data about data." It provides descriptive, structural, and contextual information, making other data easier to understand, locate, and use effectively. By capturing essential details—such as a dataset's origin, structure, purpose, relationships, and meaning—metadata enables data to be organised and contextualised.

Here is a key insight: Semantics gives data meaning. By adding semantics to the metadata, we make all metadata meaningful.

According to the Free Energy Principle, the boundary of a system is coupled to its environment in a way that selectively regulates information exchange. This selective coupling helps the system resist entropy, increasing order and minimising uncertainty. Rosen goes further, suggesting that in living systems, the presence of this boundary enables internal Closure To Efficient Causation. Efficient causation, an Aristotelian term, refers to the immediate source of change in a system—the "cause" that directly brings about an effect. In other words, only a system with a boundary can possess agency, autonomy, and self-determination.

If you’ve struggled to take your LLM project from prototype to production, and even tried RAG but still didn’t achieve the accuracy you needed, it might be time to consider GraphRAG. GraphRAG combines the power of retrieval-augmented generation with the structure of knowledge graphs, delivering more reliable and accurate results.

LLMs like ChatGPT have taken the world by storm, but for enterprises, they are only half of the equation. Knowledge Graphs (KGs) are the other half, providing the reliability and structured understanding that LLMs lack.

When discussing AI, we often focus on the algorithms—the visible 'tip of the iceberg.' But let's not forget what's submerged: a complex framework of data pipelines. The engineers who build and maintain these pipelines work tirelessly to clean and aggregate vast datasets. In fact, most of the grunt work often goes into this data preparation.

Traditional filtering is typically based on tabular (rows in a database) or tree-like (JSON documents) data formats. The landscape changes significantly when the data itself is structured as a graph. When employing HNSW in a graph-based setup, both continuous vectors and discrete facets become vertices in the same graph. This allows for more nuanced relationships and more efficient alignment. Furthermore, the upper layers within HNSW represent a form of compression. With your data in a graph, you can move beyond the classic HNSW node-degree compression algorithms to consider more semantic forms of compression, which take domain-specific ontologies into account.

Vectors need Graphs! Embedding vectors are a pivotal tool when using Generative AI. While vectors might initially seem an unlikely partner to graphs, their relationship is more intricate than it first appears.

Data leaders are adapting to the profound shift brought about by GenAI. As organizations incorporate AI into their data strategies, Graph Retrieval-Augmented Generation is emerging as a transformative solution, bridging the gap between AI and Data. This post explores GraphRAG and how it integrates into your broader data strategy.