Enterprise data was designed for people. AI agents raise the standard.
Dimensional modeling has given enterprises a durable way to organize facts, dimensions, business context and history for decades.
That foundation remains valuable.
But AI agents introduce a new kind of analytical consumer.
An analyst can recognize stale data, remember that a business definition changed, know which source is authoritative, understand that two identifiers refer to the same entity, or question a number that does not look right.
An AI agent cannot safely depend on that unwritten organizational knowledge when it is expected to recommend, decide or act.
Agentic Dimensional Modeling explores how dimensional architecture should evolve for that environment.
At the center of the book is Agent-Ready Dimensional Modeling (ARDM), an approach that strengthens the analytical contract around machine consumers.
You will learn how to design for:
- business events and machine decisions at the correct grain
- facts for events, state, recommendations, decisions, actions and outcomes
- dimensions that carry governed semantic meaning
- enterprise identity across analytical and operational systems
- current truth versus decision-time knowledge
- source authority, provenance, quality and evidence
- durable agent identity and deployment history
- policy evaluation, approval and human oversight
- attempted actions versus committed business transactions
- reconciliation between agent activity and authoritative business state
- measurable business outcomes
- boundaries between dimensional models, semantic layers, MDM, graphs, vector stores, documents, policy services and operational APIs
The book uses the continuous Atlas Industrial Group case across its 22 chapters so that the same inventory, supplier, pricing, revenue and service decisions can be followed from business requirement through modeling and production architecture.
The book is organized around four stages:
Define the Business Problem → Define the Solution → Design the Model → Deliver the Platform
It is written for practitioners who already understand relational modeling, facts and dimensions, star schemas, SQL and warehouse or lakehouse fundamentals.
A free Reader Edition companion pack provides ten ARDM workbooks with supplied inputs, practitioner exercises, worked reference solutions and design rationales.
The central idea The data warehouse taught enterprises how to organize information for decision support.
The next challenge is making that information understandable, reconstructable and governable when software itself participates in the decision.
That is the architecture developed in Agentic Dimensional Modeling.
After reading the book, practitioners should be able to:
- translate an agentic business requirement into an analytical architecture
- determine the right grain for business events, recommendations, decisions, actions and outcomes
- design dimensions that provide reusable governed meaning to AI agents
- reconstruct what the enterprise knew when a decision was made
- distinguish source authority from freshness, quality and model confidence
- represent evidence without turning dimensional facts into metadata stores
- model agents as governed analytical entities
- trace recommendation → decision → approval → action → commitment → outcome
- determine what belongs in dimensional modeling and what belongs in surrounding architecture
- evolve an existing BI-oriented warehouse or lakehouse toward an agent-ready analytical platform
