Posts tagged "data-modeling"
- From Data Vault to an AI-ready silver layer: same build order
Vault steps build an AI-ready silver layer's meaning layer: unhashed business keys, persistent staging audit, data contract, not point-in-time or bridge tables.
- You don't need to unlearn Data Vault to model for AI agents
In data modeling for AI, Ensemble Logical Modeling becomes a business ontology: the meaning layer of an AI-ready data architecture. It doesn't need the tables.
- AI for modeling, or modeling for AI?
AI-assisted data modeling guesses. Data modeling for AI writes business context for AI once as metadata in a context layer: AI proposes, a person approves.
- Who is the data model for?
A data model for AI isn't a new table shape. The model did two jobs: a diagram for people and a build instruction for machines. AI retires the second.
- Seven steps to a meaning layer
Where to start on a silver layer an AI agent can read: one business area, seven steps in order, and a check you can run at the end of each one.
- What a hub becomes when an agent reads the model
Data Vault and third normal form beside AIMS, construct by construct: where each hub, link and same-as link goes when an agent reads the model.
- Every method stops at the concept
Silver layer modeling methods say which business thing a table is about. Only the costliest says what each column means, and that's where AI answers break.
- You don't need Data Vault anymore
You don't need Data Vault or third normal form in your silver layer. AI reads business meaning from an ontology: the deciding stays, the rebuilding goes.
- Modeling for Machines That Can Read
Once a machine can read a sentence about your business, every benefit Data Vault's tables bought has a cheaper home. That's data modeling for AI.
- The economics of meaning flipped
Stated meaning now costs a sentence, not a table. What an ontology adds to AI accuracy on the benchmarks, and where the AI still has to stop.
- A Model You Can Diff: Metadata Version Control for Data Teams
Pipeline code gets code review and an audit trail; metadata doesn't. Five metadata version control practices bring naming conventions under data governance.
- Six steps to your integration layer: from bronze to trusted gold
Bronze shipped and gold still isn't trusted. Here are six steps to the integration layer your medallion architecture is missing.
- Top-down and bottom-up data modeling: cart before the horse
Top-down and bottom-up data modeling: join the conceptual data model and business glossary to source system discovery, so staging tables carry data governance.
- This is why we need data engineers
Will AI replace data engineers? We built Ask AI into DeltaVault, then recorded it losing a race to a data engineer building one view by hand.
- No silver layer, no lakehouse
A medallion architecture with no defined silver layer is a dump house, not a lakehouse. Here is how to tell which one you built.
- AI drafts the tedious part. You keep the decision.
AI-assisted data modeling in DeltaVault: skills draft keys, joins, views and descriptions as proposals you read first. Nothing writes until you accept.
- Meaning Leaks at Every Handoff
Business definitions in one tool, the technical data model in another, lineage in a third: meaning leaks at every handoff. DeltaVault keeps them in one model.
- The Medallion Three-Body Problem
Without an agreed business model, bronze, silver, and gold drift apart: three suns pulling the medallion architecture into a three-body problem.
- The Lakehouse Ended One Satellite Per Source
One satellite per source is a row-store habit. On Delta Lake width is cheap and joins are what you pay for: consolidate Data Vault satellites by meaning.
- No Human, No Loop
Human in the loop puts a person at the end of an AI process as a checkpoint. In DeltaVault the human starts the loop, so there is no loop without one.
- The Model Says Yes. The Source Says No.
Your conceptual model says delivery address. Your source says address type. Describe the view in one sentence and keep column-level lineage on the result.
- Stop Asking Data Engineers to Paint
A conceptual data model breaks on first contact with a real source system. The fix isn't better artists. It's both sides recorded as metadata.
- The Meeting Is the Model
Turn a Teams meeting transcript into a conceptual data model in one workshop run, with every proposal traceable back to what was said in the room.
- How raw is your Data Vault?
In a raw vault, raw describes the data, not the structure. Why source-shaped vaults never integrate, and how to rebuild yours without starting over.
- AI Data Modeling: Enough Talk. Let's See It in Action
The AI business modeling workshop drafts concepts, relationships, and facts from your own content, shows its reasons, and writes nothing until you apply.
- How to stop the AI Guessing
The AI business modeling workshop checks your stories before it proposes a single concept. Here are the rules it applies.
- Where, Oh Where, Has Integration Gone?
An AI can model every source you own and still leave your warehouse un-integrated. Data integration is a decision about business keys, and AI can draft it.
- AI does not fix bad data models
AI data modeling aimed at your source schema returns a faster, shinier source schema. Business meaning comes from your own material, and sources map into it.
- Give AI Content, Not More Prompts
Upload, consolidate, and research business content into DeltaVault's AI knowledge base, then review every draft before it shapes your business model.
- Built, Not Painted
When the source does not hold the model's shape, AI helps build the bridge: discovered relationships, drafted views, and column-level lineage end to end.
- Business context is not an AI prompt
Business context for AI in DeltaVault is not what you pasted into the chat box. It is a governed setting every AI skill reads, written once in your own words.
- Meaning before machinery
AI made pipelines cheap to build. The scarce thing now is business context for AI: what your business means. DeltaVault captures it once and builds from it.