Thinking on semantic data modeling and the universal semantic layer, plus release notes on what's new in Ellie.
Semantic, conceptual, and logical data models are usually compared by their differences, but they share common ground: all start from understanding the business, build a shared language, capture relationships, and prevent costly problems early, so treating them as connected assets rather than one-off documentation strengthens governance and AI initiatives alike.
Turn any folder into a review workflow: edits land in a development branch, reviewers comment and approve, and master stays stable until you merge.
Conceptual models help people agree on what business terms mean; ontologies formalize that meaning so software can reason over it. Skipping the first step is why AI projects stall.
AI can already automate the technical side of data modeling, like documentation and schema analysis, but resolving ambiguous business definitions, such as what counts as a “customer” across marketing, finance and support, and building stakeholder alignment remain decisions only people can make.
Ellie 9.7 adds dedicated View and Edit canvas modes plus an entity preview sidebar, so you can review or work on a model without ever losing your place.
Action Triggers let you fire a webhook straight from a model, so creating a ticket, starting a review, or notifying your team no longer means leaving Ellie.
Why compliance in banking, insurance, and healthcare is a consistency problem rather than a reporting one, with examples from BCBS 239, Solvency II, and HIPAA showing how semantic modeling keeps concepts like “customer” and “coverage” consistent across the teams and AI systems that rely on them.
Conversational BI fails without shared meaning. An enriched data model, the semantic layer, is what lets AI give reliable, repeatable answers instead of confident guesses.
Traditional data modeling tools are built for technical documentation, not cross-functional collaboration. Here are seven reasons that gap fails business, governance, and technical stakeholders.
For banks, the data model is now the contract behind KYC, AML and risk definitions. Semantic modeling, FIBO alignment and native Collibra/Purview/Fabric integration make it governed and traceable.
Traditional data modeling tools weren't built for banking's governance, KYC, and regulatory reporting demands. Here's what modern enterprise modeling platforms need instead.
How AI is opening semantic data models to non-architects: a built-in chatbot for plain-language questions, an MCP server that lets AI agents create and edit models directly inside existing workflows, and AI-assisted metadata enrichment for reverse-engineering legacy systems.
A 45-minute call with a data modeling expert, not a sales pitch.