Six questions to ask any AI provider about your data.
The question leaders ask us most is “what happens to our data?”. Ask any AI provider these six questions before you sign.
1. Will it train on our data?
On business and enterprise plans from the main providers, training on your data is switched off by default. On free and personal plans it may be on unless someone switches it off, so never use personal accounts for work.
2. Where is our data processed and stored?
It depends on the provider, the plan and sometimes the model. Some offer UK or EU storage, but the processing may still happen elsewhere. Ask about both, and get the answer in writing.
3. How long is it kept, and can we delete it?
It ranges from days to years, depending on the plan and settings. Agree a period that matches your own retention policy, and check how deletion works.
4. Who can see it?
When an assistant is connected to your files, a well set-up one shows each person only what they can already open. If your folder permissions are messy, AI will surface the mess, so tidy them first.
5. What does the contract say?
Look for a data processing agreement, the provider’s list of sub-processors, its security certifications, and what happens to your data if you leave.
6. Do we need a private set-up?
For very sensitive data, options include running models in your own cloud account and region, or open models you host yourselves. They cost more to run, so match the set-up to how sensitive the data is.
One thing to remember
The plan you choose often matters more than the brand. Terms change as new models are released, so check them before you sign.
These six questions come from our guide, AI in plain English. We checked these points on 23/10/2026.
Not sure where to start?
A free 30-minute conversation with a senior adviser tells us both what would help most.