Building Enterprise Intelligence with Microsoft Fabric Ontology
Insights from Poonam Chug [VP – Business Unit Head] and Johan Krebbers [Chief Technology Officer], Acuvate
About This Episode
Join Poonam Chug and Johan Krebbers in Episode 13 of Coffee Conversations as they explore how ontology can help enterprises move beyond disconnected data and give business users and AI agents a shared understanding of how the business actually works.
Most organizations already have the information they need across databases, reports, documents, emails, Teams conversations and external sources. The challenge is connecting that information with the right business context — understanding how a customer relates to an order, a shipment, a supplier or a business outcome.
Poonam and Johan discuss how Microsoft Fabric Ontology, together with Fabric IQ, Foundry IQ, Work IQ and Web IQ, can create a common business context layer across these sources. This enables users to ask questions in familiar business language while giving AI agents more grounded, traceable and context-aware information to work with. Microsoft similarly positions Fabric IQ around translating enterprise data into business concepts and relationships.
Key Topics Discussed
- Entities, relationships and actions in ontology: how business concepts are modeled, not just connected.
- Reuse of existing semantic models: ontology sits on top of them rather than replacing them.
- Fabric IQ + Foundry IQ + Work IQ + Web IQ: each adds a different layer of enterprise context.
- Grounded responses with citations and traceability: users can see where an answer came from.
- What-if analysis and scenario planning: e.g. changing carriers, routes, shifts or production plans and assessing impact.
- Past, present and future context:combining historical, real-time and planning data.
- Role-based access and governance: responses respect the user’s existing Microsoft permissions.
- Knowledge graph creation: mapping business entities and relationships visually with business users.
Want to Hear the Full Conversation?
Fill out the form to watch Episode 13 and discover how Microsoft Fabric Ontology can help connect enterprise knowledge, give AI agents richer business context and make data easier for business teams to use.
Intelligent P&ID Digital Twin - FAQs
Microsoft Fabric Ontology is a business context layer that helps organize enterprise data around concepts people understand, such as customers, orders, shipments, products and suppliers. It defines how these entities relate, allowing users and AI applications to work with business meaning rather than just tables, rows and columns.
A semantic model helps organize data for reporting, metrics and analytics. Ontology builds on that foundation by adding business entities, relationships and context across different data sources. Organizations can therefore reuse existing semantic models instead of replacing them.
Ontology gives AI agents a shared understanding of business concepts and relationships. When combined with enterprise knowledge sources, agents can provide more grounded and traceable answers, including references to the documents or sections used to generate a response.
Each layer adds a different type of context. Fabric IQ connects structured and operational data, Foundry IQ brings in documents and knowledge sources, Work IQ adds workplace context such as emails and Teams conversations, and Web IQ can incorporate trusted external information. Together, they help provide a more complete view of a business question.
Yes. Once business entities and relationships are defined, users can ask questions using familiar business terminology instead of navigating several reports or depending on IT teams to retrieve the information.
Yes. The conversation describes scenarios such as evaluating a different shipping carrier or route, understanding cost impact, and assessing changes to production capacity. By combining historical, real-time and planning information, teams can investigate possible business outcomes before making decisions.
Rather than modelling the entire enterprise at once, Acuvate recommends starting with one important business problem. The team identifies the relevant entities and relationships, connects the required data sources, builds an initial working model and then expands it as value is demonstrated.