Data and analytics
Data you can rely on
We connect external sources, match and verify data, build data marts and embed an AI assistant into the client's product.
What we do
From the source to the answer
-
Connecting external sources
We connect external data sources and bring them to a common format: loading, normalization, completeness checks.
-
Matching and verification
We check data against reference directories, find duplicates and discrepancies, and link records from different sources.
-
Data marts
We build data marts for business tasks: metrics are calculated once and the same way for every report.
-
An AI assistant in the client's product
We embed an assistant into the client's product: a user asks a question in plain words and gets a report.
Principle
Numbers come from the calculation layer, they are not generated by the model
The model understands the question and chooses the right calculation. The figures themselves are computed by the calculation layer from data marts, so every answer can be checked: you can see which data it came from.
AI assistant
How a question becomes a report
-
Question
The user asks in plain words, right in the interface of their product.
-
Interpretation
The assistant identifies the metrics, breakdowns and period.
-
Calculation
The calculation layer computes the metrics from data marts within the user's access rights.
-
Report
The user gets a table or a chart with a short commentary.
Reliability
Access rights, quality and updates
-
Access rights
Users see only the data they are entitled to; the assistant works within the same boundaries.
-
Data quality
Completeness and consistency checks on every load; discrepancies are visible before they reach a report.
-
Updates
Sources are refreshed on schedule, data marts are recalculated automatically, and the load history is kept.
Have data you need to work with?
Tell us about the sources and the task, and we will suggest how to connect the data and what can be built on it.