From the day it arrives to the day it is deleted
Before any data arrives, the client approves a data agreement and a set of handling instructions. Every consultant signs the agreement before receiving the data. This is what they commit us to.
01Approved places only
The data lives in a project folder on a computer only the consultant uses, or in U-M Dropbox and U-M Google Drive. Nowhere else: no iCloud, no OneDrive, no personal accounts, no phones or USB sticks, no email, no printouts.
02Set up before downloading
Disk encryption on. Synced folders avoided. Tools that save copies in the background, and AI assistants that read files or the screen, turned off while the data is open.
03Shared by name, not by link
Teammates get files through U-M Dropbox or U-M Google Drive, added one by one. Code goes into a private repository with no data in it. Data stays out of chat apps.
04AI only where it is permitted
The data goes only into the university's own AI services. Our analysis tool, Claude Code, never sees it: the code is written on synthetic data and runs on the real records in a locked folder.
05No one is identified
We never try to work out who anyone in the records is, and we never link the records to outside information about individuals.
06Small groups are not broken out
In anything meant to travel beyond the client, a group too small to report is folded into a larger one. Every figure says which records it counts and how many.
07The client hears it first
What we find goes to the client. Anything that leaves the class is shared only if the client agrees in writing.
08Deleted at the end
When the work is done, and by the date agreed with the client, every copy of the data and of anything with findings in it is deleted, cloud trash included.
Questions about how we would handle your data?
We are glad to go through it with you before anything is shared.
Talk to us