Data security
Access to your data limited to what is strictly needed
We develop the logic of your reporting without accessing your raw data. The access level is agreed with you and your IT department at scoping, before any work starts.
Principles
- Data minimisation
- By default, we work on the code (Power Query M, DAX, Python) and control aggregates, never on your raw data.
- A level chosen with you
- Agreed in writing at kick-off, based on how sensitive your data is and your IT department's rules.
- No use beyond the engagement
- Your data serves your engagement only. When it goes through a third-party service, AI tools included, it is through enterprise accounts that exclude training, under contract (DPA).
Four levels
What leaves your environment, depending on the level
- N0Code and logic only (M, DAX, Python), control aggregatesDefault
- N1Anonymised data or synthetic samplesPrototyping, demos
- N2Real data through third-party services under enterprise contract, no trainingWith IT approval
- N3Processing deployed in your own cloud (Azure, AWS, GCP)Large accounts
Anonymisation
Support as soon as you aim for a higher level
Pseudonymisation of third parties, employees and personal data before anything goes through a third-party service, with documented rules your IT department or DPO can approve. Included in Socle, optional on an existing engagement.
Let's talk about your constraints
We present this grid at the first meeting and set the level with you and your IT department.
Request an assessment