Letting an agent operate your business raises a fair question: whose data is it, and who gets to see it? The short answer is that it stays yours. The longer answer is worth understanding, because the design is what makes the short answer true.
Your data stays yours
The business data in your workspace belongs to you. An agent operating it is doing work on your behalf, the same as an employee using your systems; that does not transfer ownership. You can export it and you can delete it, and the platform is the custodian, not the owner.
What the agent can and cannot see
An agent sees only what the person it acts for is allowed to see, enforced when tools are offered and again when they run. It is not a window into everything; it is scoped to a role, exactly like a human user. Bring-your-own-agent does the reasoning at your chosen provider, and operates your workspace through that permissioned boundary.
Isolation between businesses
Each workspace is isolated, and on a branded network each customer's data lives separately and is never shared across tenants. One business's agent cannot reach another's data, because the boundary is structural, not a matter of politeness.
Training and your content
Your business data is not someone else's training set. When you bring your own agent, the model is whatever you chose, under that provider's terms, and your operational data is used to do your work, not to train a model behind your back.
Your rights, in practice
Ownership only means something if you can act on it: see what is held, export it, and have it deleted on request. Those rights are built in, with a clear route to make a data request, so control is not just a promise but a button.
In the agent era, the data is still yours; the agent just does the work on it. See how isolation and permissions work, read the privacy terms, or launch a workspace.