AI on your data
A language model that doesn't know your data makes answers up. We connect Gemini to your data warehouse through the same semantic layer as your dashboards. Ask what the margin on new customers was last month, and the AI uses your definition of margin and of a new customer.
What we often hear
- We want AI on our data, but don't know if the answers are right
- ChatGPT makes up numbers when we ask about our figures
- Not everyone can build a report, but everyone has questions
What we do
Gemini on your warehouse
Through Vertex AI in your own Google Cloud project, so your data stays in your own environment.
Your definitions
The AI uses the definitions from your semantic layer, not its own interpretation.
Limits
Questions outside your data, or outside the user's permissions, the AI doesn't answer.
Test questions
A fixed set of questions we use to check that the answers are right, and that your team can start with.
What we work with
The tools we use for this. Which ones we pick depends on your setup.
Frequently asked questions
Do I need a data warehouse?
Yes. The answers are only as good as the data and definitions underneath. If you don't have a warehouse yet, we build it. If you already have one with a semantic layer, we build the AI on top of it.
What does using it cost?
You pay Google directly for Gemini, through your own project. How much depends on how many questions your team asks. We give you an estimate up front.
How to buy it
We work in three steps: first a plan, then the build, then maintenance.
Blueprint
Which data sources, model and setup fit the questions you have.
Build
Your data in one place, with dashboards and an AI layer as options.
Maintenance & growth
Pipelines and models stay correct when your data sources change.
Want to know what this means for your situation?