If your company runs on Microsoft 365, AI buyers can license a de-identified copy of your Teams chats, Outlook mail and SharePoint and OneDrive files, priced as one connected record. The largest public example is Spirit Airlines' archive of 100 million emails and 500 million Teams messages, which Google won with a $10 million auction bid, a sale the court has not yet approved (The Next Web, August 2026; TravelPulse, October 2026). For small and mid-size companies, buyers publish ranges that mostly start near $100,000.
On this page
Why do AI labs want Teams, Outlook and SharePoint records?
AI companies are training agents to do office work inside tools like these. Microsoft 365 holds the three things they need side by side: the conversation (Teams and Outlook), the work product (Word, Excel and PowerPoint files in SharePoint and OneDrive) and the timing that links them. A Teams thread about a late order, the email to the customer and the revised spreadsheet together show how a team solved a problem.
Buyers name these tools directly. Scale AI lists Teams for chat and Word, Excel, Slides, SharePoint and Drive for files, and asks for file contents "plus their metadata, authorship, timestamps, and access permissions" (Scale AI, October 2026). G2i lists Outlook, SharePoint, OneDrive, Word, Excel and Teams, and says one app is enough to start but connected apps are worth more (G2i, October 2026). Others that list Microsoft 365 records include Polyshares, Troveo, Grepped and SellMyData. See Scale AI and G2i for their terms.
Data Licensors, the company that publishes this page, also licenses company records from Teams, Outlook, SharePoint and OneDrive. Our buyer page sits next to the others.
What is Microsoft 365 data worth?
| Reference point | Amount | Kind of figure | Source |
|---|---|---|---|
| Spirit Airlines archive: 100M emails across 80,000 accounts, 500M Teams messages, 17.1M OneDrive files, 20.6M SharePoint files | $10,000,000 winning bid by Google | Auction bid, not final: hearing set for November 6, 2026 | The Next Web, August 2026; TravelPulse, October 2026 |
| Same archive, Mercor's next-highest bid | $7,500,000 | Auction bid, backup bidder | Axios, August 2026 |
| Same archive, Mercor's offer if it could take the raw data and de-identify it itself | $7,000,000 | Auction offer, not chosen | The Next Web, August 2026, citing Business Insider |
| Same archive, micro1's late offer | $12,500,000 | Offer after the auction; the court is still weighing Google's bid | Fortune, September 2026 |
| micro1 published tiers for operating companies | $100K to $500K, $500K to $2M, $2M+ | Published range | micro1, October 2026 |
| Handshake AI published range | $100K to $4M+ | Published range | Handshake, October 2026 |
| Broker table for 25 to 100 staff (workspace, email and support archives) | $100K to $500K | Broker claim, from "one advisory firm's published table" | Your Data Value, October 2026 |
The airline is not a guide for a 40-person company: about 17,000 people lost their jobs when it stopped flying, and its employee records go back to 1986 (The Next Web, August 2026). It is useful for one lesson about de-identification. The sale agreement lets the buyer choose and pay the firm that removes personal details, and requires the cleaned data to stay linked across email, tickets and code (The Next Web, August 2026). Google told Business Insider (October 2026) that any personal information would be excluded or de-identified by an independent third party. Before you sign with any buyer, ask who de-identifies your records, who pays for it, and whether you see the result first. See all reported deals in our deal ledger.
How do I check how much Microsoft 365 data I have?
All of these are in the Microsoft 365 admin center under Reports, Usage, and each one exports to CSV.
| What you want to know | Report | Field to read |
|---|---|---|
| Teams messages | Teams user activity | Channel messages, Chat messages, Posts, Replies |
| Mailbox size and total mail | Exchange mailbox usage | Item count, Storage used |
| Email sent and received | Exchange email activity | Send actions, Receive actions |
| SharePoint files | SharePoint site usage | Files, Active files, per site |
| OneDrive files | OneDrive usage | Files, per account |
Source: Microsoft Learn, checked October 2026.
The activity reports cover the last 7, 30, 90 or 180 days at most, so double the 180-day Teams count for a yearly rate and multiply by the years you have used Teams for a rough total. The file and item counts are totals as stored today. Note the year your tenant started and your peak headcount; buyers ask for both.
How do I export Microsoft 365 data safely?
Microsoft Purview eDiscovery is Microsoft's built-in way to search and export mailboxes, Teams conversations, SharePoint and OneDrive together. Points from Microsoft's documentation, checked October 2026:
- Mailbox items export as PST files or single messages. SharePoint and OneDrive content exports as native Office and other files.
- The user who runs the export needs a Microsoft 365 E3 or E5 license, and a SharePoint E3 license for SharePoint content.
- You choose document versions: latest only, recent 10, recent 100, or all.
- An option includes Teams messages posted 12 hours before and after each matching item, which keeps conversations whole.
- Files restricted by data loss prevention rules, or in locked sites, are skipped.
- Export packages expire 14 days after creation, and an export that runs over 7 days is canceled. Split by date range or by team.
If your plan does not include eDiscovery, smaller exports are still possible. Each user can export a mailbox to a .pst file in Outlook for Windows. Classic Outlook includes mail, calendar and contacts; the new Outlook (Settings, Files, Export) exports mail only (Microsoft Support). SharePoint document libraries can be synced or downloaded by a site owner.
Keep every export on encrypted company storage, send it only through the encrypted transfer a buyer gives you, and delete working copies when delivery is done. At Data Licensors your team runs the export with our step-by-step guidance, or grants a limited read-only connection you can revoke. We never ask for passwords.
What should I leave out?
You set the scope by team, site, mailbox and date. Common exclusions:
- HR, payroll and finance leadership sites and mailboxes.
- Personal OneDrive folders. People keep personal files there. Licensing shared team sites is cleaner.
- Teams private chats, unless the people in them agree. Channel conversations hold most of the process value.
- Shared channels and guest content from other organizations. Their messages are not yours to license.
- Files under client NDAs or with sensitivity labels your contracts attach to client data.
- Mail with your lawyers, to protect privilege.
- Health information, payment card data and government ID numbers. We never accept them.
How is Microsoft 365 data de-identified?
Each part of Microsoft 365 hides identifiers in a different place:
- Teams: display names, @mentions, team and channel names that include client names.
- Outlook: headers, signatures, quoted replies and calendar invites with attendee lists.
- SharePoint and OneDrive: file names, document text, comments, tracked changes and the author and "last modified by" fields in file metadata.
Names, emails, phone numbers, addresses, company and customer names, and account numbers become placeholders such as [Name] or [Customer]. Secrets such as passwords in spreadsheets are removed. Timestamps, version history and the links between a thread and a file stay. A person reviews samples from every batch, and you see a de-identified sample before you sign. More detail: how de-identification works.
What should I do next?
Pull the five reports above, decide which teams and sites are in scope, and list what stays out. Then get a quick estimate and ask more than one buyer to price the same scope in writing. If you also use Gmail or Slack, see our guides on email archives and Slack data. Before signing, check the terms in our data licensing agreement guide.
