Work-managed accounts require confirmation of who controls the settings; personal accounts require a direct review of those settings. In both cases, the reader should verify model-training controls, chat history, data exports, and account deletion. The central distinction is that turning off model improvement means new conversations are not used to train models, but they can still appear in chat history.
Start by identifying who controls the account
For a work-managed account, confirm who administers it and whether its settings can be viewed or changed directly. The reader should not assume that every control available on a personal account is also available on an organisation-managed account.
For a personal account, confirm that the intended account is being used and review its current settings directly. The account type alone does not establish how data controls are configured.
What to check in each account type
| Verification point | Work-managed account | Personal account |
|---|---|---|
| Control access | Confirm who can view or change privacy and data controls. | Review which controls are available and whether they can be changed directly. |
| Model improvement | Ask the administrator whether the model-improvement option is available and whether it is off. | Check the current model-improvement setting. Turning it off does not mean new conversations will be absent from chat history. |
| Chat history | Confirm whether new conversations can remain visible and who can access or manage the history. | Check chat history after reviewing the setting: “not used for training” and “not present in history” are different conditions. |
| Data exports | Confirm whether an export function is available and who can use it. | Check whether data export appears in the current account settings. |
| Account deletion | Confirm what deletion process is available and whether account-specific conditions apply. | Review the current deletion option and its stated conditions before relying on it. |
The distinction that matters most
The relevant help documentation identifies controls for model training, chat history, data exports, account deletion, and other privacy options. It also explains that turning off the model-improvement option prevents new conversations from being used to train models while allowing those conversations to appear in chat history.
That setting therefore answers a model-training question; it does not by itself answer a chat-history question.
What the reader must still confirm
The cited documentation does not set out how work-managed settings differ from personal-account settings. It also does not specify administrator permissions, retention periods, export limits, or deletion conditions. Those details remain account-specific and must be confirmed through the current account interface or, for a work-managed account, the relevant administrator.