Depositori

About us

Why we built Depositori.

Who makes it, and what we believe about the numbers a business runs on.

Background

Data people who kept being asked what a number means.

Depositori is made at In2Intel, a data tech agency in Amsterdam. Our team includes data engineers, scientists, and for years we have built data platforms, models and BI reports for clients across multiple sectors includings retailers, hospitality, and firms in real estate and private equity.

Between us we cover the whole way a number travels: the integrations with ERP, CRM and POS systems, the pipelines and models in the warehouse, the reports on top, and lately the AI agents that answer questions from company data: Depositori grew out of that work.

Few companies run on one clean stack. Most have tools that grew up side by side over the years. Depositori is built for that: a hub with spokes, where each system plugs in on its own and the glossary in the middle ties their definitions together. No rebuilding required.

It holds when the stack changes, too. The definitions live in Depositori, not inside one tool, so a move to another platform takes them along instead of rewriting them by hand. They are kept in an open standard for semantic models (Apache Ossie), so the work you do in Depositori goes with you, even if you leave Depositori itself.

Manifesto

  1. 01

    A number is only as good as its definition. Two reports can both be right and still disagree, because they count different things under the same name.

  2. 02

    Definitions belong to the business. The people who use a term should be able to read it, question it and approve it, without reading code.

  3. 03

    Business and tech work side by side. The business decides what a term means, engineers decide how it is built. Each works on their own part, sees the other's, and nobody waits on anyone.

  4. 04

    Change a definition once. When the team agrees on a new one, every report, model and agent should follow, not only the ones someone remembered to update.

  5. 05

    A migration doesn't have to cost a fortune. When the definitions live apart from the tools, moving to a new platform means carrying them over and checking the numbers still match, not rebuilding every report by hand.

  6. 06

    AI on your data is possible today. It takes no new platform and no year of preparation. Give an agent the definitions your team already agreed on, and it can answer from the data you already have.

  7. 07

    Agents need the same ground truth as people. An AI that answers from your data should use the definitions your team agreed on, and say which one it used.

  8. 08

    Your data stays where it is. Knowing what a number means should never require copying the rows behind it.

  9. 09

    Built on open standards, so you're not locked in. Your definitions are kept in an open format. AI models can read them as they are, and you can take them to another tool whenever you want.