Every team has the numbers. None of them agree.
Semcore gives the words your business runs on one agreed meaning, and hands the same answer to every report, application and assistant that asks for it.
Who this is for
One problem, and it does not look the same from everywhere
The person who signs the figure and the engineer building on it are living with the same disagreement and would not describe it the same way. Each card opens the page written in the reader's own words.
- AI engineersYour assistant infers what a table means.It gets it slightly wrong.
- B2B SaaS companiesYour customers read your numbers.Every surface derives them separately.
- Data practitionersYou can say what the number means.Nothing downstream can.
- Design partnersSoftware gets shaped by whoever is in the room.This is the invitation.
- EnterprisesSomebody signs the figure.Nobody owns what it means.
- Small and medium businessesOne person knows what the numbers mean.That is the risk.
- StartupsYour estate is small.The cost of settling what it means grows with it.
The problem
Nobody is wrong. There was never one place to be right.
A metric means one thing in the warehouse, another in the dashboard, and a third in the model that shipped last week. Each definition was reasonable where it was written. None of them was written anywhere the others could see it.
Meaning lives with whoever wrote it
The definition of an active customer sits in a query, a transformation model, a spreadsheet and somebody's head. All four have drifted.
A change arrives as a number that moved
Nothing announces that a definition changed. It surfaces when a figure in a board pack stops matching the one in the report.
Every AI consumer starts again
Each new assistant works out what your business words mean from raw schemas, and each one gets it slightly differently.
And the stack is consolidating. Ingestion, transformation and semantics arrive inside single vendors, which is lock-in at the layer you can least afford to rent: the one that decides what your data means.
The answer
Agree it once, and everything reads the same answer
Semcore holds what your terms mean above the systems that store your data and beneath everything that asks questions of it. A definition is agreed once, governed the way code is governed, and served to whatever needs it. A layer that does that is a semantic core, and this one is yours to take elsewhere.
Where people start
Outcomes, and the catalog entries behind them
Each entry below stands behind one of the outcomes Semcore leads with, and names the one it stands behind. Follow it for what that outcome rests on, which is sometimes a pair: an interface with nothing behind it is not something anyone can buy.
- 01Agentic Modelling CopilotDomains, models and definitions drafted across your catalog, rather than assembled by hand.Modelling that consumes analyst quarters arrives as a draft to review.Modelling where you work
- 02Semcore MCP ServerThe semantic core reachable from the assistant already in your editor, without leaving it to find out what a table or a metric means.Governed definitions are used in the tool engineers already live in, rather than consulted somewhere else.Modelling where you work
- 03Model Change ReviewEvery change to a data model put in front of a reviewer before it lands, the way code already is.A change to what a number means is approved and recorded, and nobody learns about it from a figure that moved.Review before it ships
Built to bind to the data catalogs, warehouses and stores you already run
An AI Readiness Review
A scoping conversation, and a look at your data estate as it stands: what your AI plans will need from it, where the gaps are, and which of them to close first. The findings are yours either way.