ARTIFICIAL INTELLIGENCE · ENTERPRISE KNOWLEDGE · INFORMATION ARCHITECTURE
AI is changing how organisations create, interpret and use information. But intelligence without context is only a more convincing way of getting things wrong.
The interesting question is no longer simply what a model can do. It is whether an organisation can give that model the meaning, context and boundaries it needs to do something useful. What does a product actually represent? Which customer relationship matters? When are two records describing the same thing? Who is entitled to make that decision, and what evidence should support it?
These are not new questions. They sit at the heart of enterprise architecture, master data management and information governance. AI makes them more visible, more immediate and, in many cases, more consequential.
My interest is in that intersection: how we build organisations that can use AI intelligently because they understand their own business, information and decisions.
The foundations still matter
AI can interpret unstructured information, recognise patterns, suggest relationships and work with ambiguity in ways that traditional rule-based systems struggle to match. It can reduce the amount of manual effort required to classify, enrich, match and maintain information.
That is a significant shift. It does not, however, make business definitions agree with one another. It does not establish ownership, resolve conflicting policies or decide which representation of a customer, supplier or product the organisation should rely on. A model can recommend an answer. The organisation still needs to understand what makes that answer acceptable.
I see AI not as a replacement for master data management or architecture, but as a change in how those disciplines operate. The opportunity is to move from data clean-up towards better information creation, from static rules towards assisted decisions, and from disconnected repositories towards a more coherent understanding of the enterprise.
Four areas I explore
1. Enterprise meaning
Semantic models, canonical models, ontologies and knowledge graphs. How do we describe the concepts, relationships and rules that make enterprise information meaningful? How do we give AI more than a collection of tables, documents and labels to work with?
2. Master data in the AI era
Master data is not simply a set of records to be cleaned and reconciled. It is the result of business processes, decisions and responsibilities. AI creates new ways to identify, validate and enrich information, but the business still has to define what is important and who is accountable. The focus should move upstream, towards the point where information is created.
3. Governance, trust and evidence
Confidence is not the same as correctness. Trust requires provenance, traceability, clear decision rights and appropriate human oversight. I explore how organisations can use AI without losing sight of accountability, and how feedback from real decisions can improve the systems over time.
4. The enterprise operating model
AI changes processes, roles and the boundaries between people and systems. The challenge is not to add a chatbot to an existing process and call it transformation. It is to rethink how work gets done, where decisions belong and how intelligent capabilities become part of the operating model.
From data management to enterprise intelligence
I am developing these ideas through a practitioner-led exploration of AI and master data. It is not a catalogue of tools or a set of predictions about what might happen one day. It is an attempt to understand what is changing in real organisations, what remains stubbornly difficult, and which architectural foundations will matter as AI becomes part of everyday business operations.
The aim is not to make everything intelligent. It is to make the organisation more capable of understanding its own information, making better decisions and learning from the consequences.
Selected reading
Why AI Turns Product Data Quality from a Clean-up into a Creation Habit
Moving quality upstream, from periodic correction to a habit embedded in the creation process.
Data Architecture Patterns: Decisions for the AI Era
Choosing architecture patterns around the decisions and capabilities an organisation needs.
Governing AI-Assisted Software Engineering: Guardrails, Not a Ban
Practical governance for AI-enabled development without pretending prohibition is a strategy.
AI Governance: A Structural Perspective for Practitioners
Why responsibility and structure need to evolve alongside adoption.
The Layer Has a Name: Ontology
Moving beyond tables and labels towards an explicit model of business concepts and relationships.
Golden Records Are Beliefs, Not Mirrors
Why AI-assisted matching makes the question of what an organisation accepts as truth more important, not less.
The principle
Make complexity understandable. Make meaning explicit. Keep accountability close to the decisions. That is the foundation for using AI well, and the thread connecting the articles collected here.
Explore the Master Data perspective or browse the full article archive.