Business agrees on one thing. The architect draws a second. Production ends up with a third.
Anyone who has been inside a data platform project recognizes this. It isn’t that one particular project failed — it’s that requirements, modeling, and implementation live in different tools, moving at different speeds, owned by different people.
The problem is well known, and tools have existed to address it for twenty years or more. We’ve used them ourselves, and they do their job. They were just built for a time when data modeling was one person’s desktop application — not something business, architects, and data engineers do together.
That’s why we built Model One. Today, we’re launching it.
One model, viewed at every level
Model One is a data modeling platform where business terminology, data flows, and the physical database model live in the same workspace. It’s one model, and everyone looks at it from their own vantage point: the Logical model for shared business concepts, Flow for how data actually moves from source systems into that model, and the Physical model for the schema it compiles down to.
When “customer” means the same thing in sales and in billing, that’s one concept in the model — not two table names that happen to look similar. Once the model is agreed, it compiles straight into a production-ready schema, with nothing redrawn by hand along the way.
OneAssistant: suggests, never decides
Model One ships with OneAssistant, a built-in AI assistant that reads your actual project. It drafts entities from a plain-language brief, suggests missing attributes and naming fixes, and answers questions about the model in plain language.
It suggests — it doesn’t decide. Nothing is applied silently: every proposed change is presented for review, and a person decides what actually lands in the model.
The promise from our last post, kept
We said it before, and it wasn’t just wordplay: Model One runs inside your own Azure tenant, and your data never leaves that environment.
Try it before you take our word for it
We promised to let others in too, so now you can try before you believe. The Model One Challenge is six short, browser-based visual puzzles — five that each teach one part of the platform, and a final one where you inherit a sales order model nobody trusts anymore: duplicates, missing relationships, a data flow that goes nowhere. Your job is to fix it. No sign-up, no form — it runs entirely in your browser, and nothing you do leaves your device.
Ready to keep going with a model of your own afterward? The free Community edition is the lightest way to start: one project, up to twenty entities, and the full modeling stack — logical model, Flow, physical model, shared dictionary, and real-time collaboration included.
Want to talk it through first? Get in touch about a trial.