What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
Your business data is processed to generate an AI assistant’s response. Depending on the account, settings and features involved, the service may also log or retain the conversation, make records available to workplace administrators, or send a query to a separate web-search provider. “Not used to train models” does not mean “never stored or accessed.”
The answer depends on the exact service and plan—not simply whether the tool is called a business AI assistant. OpenAI, Microsoft and Google describe different protections and controls for different account types. Here is how to understand the data path and what to check before employees use an assistant with company information.
As an Amazon Associate I earn from qualifying purchases.
What happens to data in an AI assistant?
When you submit a prompt, the service processes its contents and relevant context to produce a response. That context might include text you pasted, an attachment, information available through a connected work account, or a web search triggered by the prompt. The provider’s terms and configuration determine what happens beyond that immediate processing.
Recommended Free Tools
Four questions are often conflated:
- Processing: Is the information handled to generate the answer or operate the feature?
- Model training: Is it used to train or improve the provider’s foundation models?
- Retention: Are prompts, responses or related records stored, and for how long?
- Access: Can administrators, auditors or other authorized people view or search those records?
A service can answer “no” to model training while still logging interactions for organizational retention, auditing or service operation. For example, Microsoft says work- or school-account Copilot Chat prompts and responses are not used to train foundation models, while also explaining that interactions are logged and may be available to IT administrators through search and audit tools (Microsoft Copilot Chat overview).
#1 Best Overall
How business data handling differs by provider and account
These examples describe the cited provider documentation, not a complete comparison of every AI service. Eligibility, features and administrative controls depend on the exact plan and can change; verify the terms for the account your organization uses.
OpenAI: personal workspaces versus business offerings
OpenAI says inputs and outputs from ChatGPT Business, Enterprise, Edu, ChatGPT for Healthcare, ChatGPT for Teachers and its API platform are excluded from model training by default. Its business documentation also describes encryption in transit and at rest, access management, retention controls for qualifying organizations, and data-residency options for eligible services. Which controls apply depends on the plan and service (OpenAI business data privacy).
That default differs from personal Free, Plus and Pro workspaces, where data sharing is on by default. A user can turn it off for future conversations in the data controls settings. This is a training choice, not an answer to how long a conversation is retained or who may access it. OpenAI distinguishes these personal settings from business and API offerings, whose inputs and outputs are not used for training by default (OpenAI Data Controls FAQ).
Rank #2
In ChatGPT Business, members have separate chat histories; they do not automatically see one another’s chats. A shared link is a deliberate way to share a conversation. Workspace spend metrics do not automatically reveal members’ full private chat histories (OpenAI ChatGPT Business workspace billing and controls).
Microsoft: Copilot Chat and Microsoft 365 Copilot
For Copilot Chat signed in with a work or school account, Microsoft says prompts and responses are not used to train foundation models and interaction data is encrypted during the chat session. It also says prompts, responses and Bing queries triggered by prompts are logged; IT administrators may use Microsoft search and audit tools to view logged information. A web query follows a separate path: Microsoft describes Bing as an independent controller with different handling terms (Microsoft Copilot Chat overview).
Microsoft 365 Copilot interaction history includes prompts and responses and is stored in line with the organization’s Microsoft 365 contractual commitments. Microsoft says this content is encrypted at rest, is not used to train foundation large language models, and can be searched or governed using Microsoft Purview (Microsoft 365 Copilot privacy).
Microsoft’s enterprise data-protection documentation says prompts, responses and Microsoft Graph data are not used to train foundation models. Copilot can inherit identity, permissions, sensitivity labels, retention policies and audit settings, although the exact controls vary by subscription. Agents can have their own terms and privacy statements, so do not assume the main Copilot protections automatically cover every agent. Microsoft also says Anthropic models are currently excluded from the EU Data Boundary when applicable; organizations with strict location requirements should check the current scope for their chosen model and account (Microsoft enterprise data protection).
Google Workspace with Gemini
Google says qualifying Workspace business and enterprise users receive enterprise-grade protections in the Gemini app: their submissions are not used to train models or reviewed by humans, interactions stay within the organization, and applicable Workspace protections such as data-region policies and data loss prevention (DLP) apply. Consumer Gemini use without a qualifying Workspace edition is different: consumer terms apply, and chats may be reviewed and used for product improvement. Confirm the organization’s edition and the terms for the specific Gemini service (Google Workspace Gemini privacy and protections).
Workspace administrators can configure Gemini conversation history. Google’s admin documentation says history is on by default and offers retention settings of 3, 18 or 36 months. If history is turned off, existing chats may remain in user accounts for up to 72 hours for service provision and feedback processing. Those are provider-described settings and service periods, not a guarantee that every account is configured the same way (Google Workspace Admin Help: Gemini conversation history).
Gemini’s access to Workspace information is bounded by the user’s own access. Administrators may limit access to Gemini or Workspace data, and content owners’ sharing settings still apply (Google Workspace Gemini access controls).
Can your employer see what you enter?
Possibly, depending on the service, account and organizational settings. Microsoft explicitly says that work- or school-account Copilot Chat interactions are logged and that IT administrators can use search and audit tools to view logged information. Microsoft 365 Copilot history can also be searched or governed with Purview. For OpenAI ChatGPT Business, members do not automatically see each other’s chat histories, but the cited workspace information does not make that a blanket statement about all administrative access. Check the controls and contract for your organization’s exact plan.
The Tool Desk
Outbyte Driver Updater FREEFix the driver behind crashes, sound loss and screen glitchesFind Drivers →Outbyte PC Repair FREERepair Windows errors before they cause bigger problemsFix Now →Being signed into a work account is not the same as making every conversation visible to every coworker. Conversely, a private-looking chat window does not prove that no administrative record exists. Ask who can search, review, export or delete interaction records, and under what policy.
Independent reader supportYour contribution helps us test, update, and keep practical guides available for everyone.What connected features change
AI assistants may draw on work files, shared drives, search, connectors or agents. Each expands the data path beyond the words typed into the prompt.
- Web search: Microsoft says Bing queries triggered by Copilot Chat prompts are handled separately from Microsoft 365 prompts and responses, under different terms. Consider what part of a prompt might be sent as a search query.
- Work content: Gemini and Microsoft Copilot can respect existing identity and access controls, but that does not repair overly broad permissions. If an employee can access a sensitive file, an assistant that uses that employee’s context may be able to use it too.
- Agents and integrations: Microsoft warns that individual agents may have their own terms and privacy statements. Verify the relevant terms for each agent or third-party connection rather than assuming the main assistant’s protections cover it.
How to assess an assistant before using it for company data
- Confirm the account and plan. Make sure employees are signed into the organization’s approved business account, not a personal workspace. Identify the precise edition and feature being used.
- Read the terms for that service. Check the plan terms, data-processing addendum and applicable retention policy. Look for separate treatment of prompts, attachments, retrieved files, responses and feedback.
- Map retention and access. Find out what is logged, how long records are kept, and who can view, search, audit, export or delete them. Confirm the administrator settings actually in force.
- Review permissions and sharing. Check the assistant’s access to files, shared drives, connectors and agents, along with the underlying permissions and content-sharing rules. Apply DLP and other controls where available.
- Check separate data paths. Determine whether web search or a third-party integration sends information under different terms, and verify the relevant processing geography if location matters.
- Apply the organization’s data rules. Do not enter information that company policy, a client commitment or another obligation prohibits from being sent to the selected service.
Why “private” is not a complete answer
Privacy depends on several distinct controls: the plan and contract, training settings, retention, administrator access, data location, identity and permissions, and the terms for connected services. A training exclusion answers only whether content is used to train models under the stated terms. It does not by itself establish that content is never stored, that no administrator can access records, or that every connected feature follows the same rules.
Quick Recap
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
Do these 3 things before closing this tab:
1Scan for outdated or missing drivers - takes under a minute2Repair Windows errors before they cause bigger problems3Fix the driver behind crashes, sound loss and screen glitches
