How do you export Outlook and Microsoft 365 email for an AI data license?
Email is where a company writes down what it is thinking, and Microsoft 365 keeps it in more places than most people expect. Mailboxes, archives, shared inboxes and Teams chats each have their own route out. The first job is learning what survives, because retention settings may already have deleted some of it.
What a Microsoft 365 tenant usually holds
| Mailboxes | Each person’s inbox, sent mail and folders. The richest source and the most personal. |
| Shared mailboxes | Addresses such as support or sales that several people read. Often cleaner than personal inboxes because they hold customer conversations. |
| Archives and older mail | Exported files from past years that people kept on their own machines. |
| Teams chats and channels | Group and one-to-one conversations. Stored in a different place from mail. |
| Calendars and meeting notes | Who met whom and when, with whatever notes were attached. |
How mail and chats leave Microsoft 365, route by route
| Tool | What an export gives you | What is missing or needs an admin |
|---|---|---|
| Outlook export by the user | A file of one person’s mail, made by that person. | It depends on each person doing it, and it leaves out anything already deleted. |
| Administrator search and export | A search across mailboxes, run by an administrator with the right role, exported to files. | It needs a role most staff do not have, and it follows the retention rules in force. |
| Teams message export | Chat and channel messages, through the same administrator search. | Edited or deleted messages, files shared in chat and anything past the retention limit. |
| The programming interface | A repeatable pull of mail and chats by a developer or partner with approved access. | It needs an application approved by an administrator. |
Tool names belong to their owners and are used only to say where records usually live. Export features change by plan and version, so check the current rules for your account.
Retention may have decided for you
Many companies set rules that delete mail after a period. Find out what yours is before promising anything about the years available. Some retention rules are legal holds, which should not be touched at all.
Shared inboxes and customer-facing mailboxes are usually the best place to begin, because the conversations in them are with customers, in the company’s voice.
Email carries more private information than anything else
Every message holds names, addresses, attachments and sometimes things nobody meant to write. Personal mail sent from a work account is in there too. Preparing a set means cleaning each of these, replacing names with stable stand-ins so a thread still makes sense.
Staff should be told what is happening, and where privacy rules require it, asked. The decision to include a mailbox is the company’s, with its counsel.
What an email set leaves out
| Personal and private messages | Excluded. Mail that is not about the company’s work does not belong in a set. |
| Names, addresses and attachments | Names and addresses are swapped for stand-ins, and attachments are screened before they are included. |
| Privileged and regulated material | Messages with lawyers, health or financial detail are left out unless they can be fully cleaned. |
An example, invented for illustration
A professional services company uses Microsoft 365 and keeps shared mailboxes for support and sales. An administrator runs a search on those two mailboxes and on a set of project channels in Teams. A set replaces customer and staff names with stand-ins, removes private messages and leaves out anything under a legal hold. The owner looks over a sample of the cleaned threads before the files are released.
Questions owners ask about this
Do we need every employee’s inbox?
No. Shared and customer-facing mailboxes are usually the better starting point, and personal inboxes are included only where staff and counsel agree.
Our mail is deleted after a period. Is it gone?
What was deleted by policy is generally gone, though some may sit in backups or archives. Ask your administrator what the rule is and what is held.
Who can run the administrator export?
A person with the right role in the tenant, often your IT lead or an outside provider. Staff without it cannot, and the search itself is logged.
Didn't find your question? Ask us directly.
Keep reading
Related kinds of records
- Support tickets: The whole thread, wrong guesses included, with customers replaced by stand-ins.
- Team chat: Slack and Teams threads, where one person must stay one person and direct messages stay out.
- Client deliverables and drafts: Each draft beside its final version with the comments that changed it, kept only where the client agrees.