What happens to my data when Charyl uses a cloud AI?

Charyl’s chat and her product research use a cloud AI provider. When they do, your request passes through us to a named provider that’s contractually bound not to train on it.

Because every request goes through us rather than straight from your home, we send it without your account — what the provider holds is a request, not a customer record. Whatever the request carries goes with it, though. To answer a question about your insurance, say, Charyl may pass along a passage from your declarations page, and that can include a name or address. A record you mark private keeps its contents out of this path, unless you open it up to her.

Apart from floor plans, which come next, we don’t keep what Charyl sends to our AI provider. (A /bug report you choose to send is separate: you see what’s in it before it goes.) The provider may hold a request for a limited time under its own policy, on a retention window measured in weeks, not forever.

Floor plans are different. When you import one, the page is sent to Charyl Cloud to be read, on these same terms, and we keep it there for up to 30 days. Floor plans usually print the owner’s name and street address in the title block, and we don’t remove it yet, so that goes along to be read too. After 30 days we delete our copy of the page — your box keeps its own — and what was read from it, mainly the room layout, stays with your import in Charyl Cloud, along with the file’s name. A plan you save to the Charyl Cloud plan library stays there until it’s deleted. You can ask us to delete what we hold; our privacy policy says how.

We don’t just take a vendor’s word for this. Before we’ll route your data through any AI provider, we hold them to a written privacy checklist first.

Which provider is behind a given answer isn’t named here — see which AI company does Charyl use?