How to find and export your SharePoint data for AI training

01

What is SharePoint data, and why do AI labs and data vendors buy it?

What a SharePoint site records

SharePoint is Microsoft's service for team websites where a company stores and shares its documents, and it comes with most Microsoft 365 business subscriptions. Each website is called a site, and every team created in Microsoft Teams also gets a SharePoint site behind it. Inside a site, a document library holds the files in folders, and a list is a table that tracks items such as requests, issues or inventory.

Sites are usually organized by department, client or project. That means a file carries a useful label from where it lives, such as a contract kept in a client's site.

Why controlled documents teach an AI model something

An AI model learns by studying a very large number of examples and finding the patterns in them, and those examples are called training data. A company's controlled documents are the examples of how its work is meant to be done: the procedure for handling a refund, the checklist for a safety inspection, the template for a contract. Public text rarely includes documents of this kind, because companies keep them inside.

A version history is the record SharePoint keeps of every earlier save of a file. It shows a document moving from a first draft through edits to an approved final, which is the process a lab wants an AI model to understand. Lists on the same sites link structured records, such as a log of customer requests, to the documents that support them. Why AI companies want your data covers the demand in more detail.

How SharePoint fits into a dataset

SharePoint files are priced as one part of a whole dataset, not on their own. micro1, the data vendor we typically refer companies to, pays them $100K to $2M or more per dataset, and the figure depends on the years of history and how completely the records show each job.

Files that belong to one person sit in OneDrive, and the conversations about shared files sit in Microsoft Teams. The Microsoft 365 overview shows how these fit together, and what your data is worth explains how the pieces add up.

02

Where does SharePoint show how much data you have?

How to read your SharePoint storage

The steps below read the size of the SharePoint footprint from Microsoft's own screens. Microsoft's guide to site storage limits and its SharePoint storage report guide describe the same screens.

1. Sign in as a SharePoint Administrator or Global Administrator.
2. In the SharePoint admin center, open Active sites. The storage bar at the upper right shows the storage still available and the total for your subscription, and the difference between the two is what you use.
3. For a tenant-wide figure and its trend, switch to the Microsoft 365 admin center and choose Reports > Usage, then SharePoint, then the Storage tab. Its Total usage chart plots storage used against your quota and the space left.
4. For detail on each site, open the Site usage tab of the same report and choose Export. Add up the Storage used (MB) column, then divide by 1,024 to get GB.
5. When site addresses show as hidden, only a Global Administrator can reveal them: under Settings > Org Settings > Services > Reports, untick Conceal user, group, and site names in all reports.

Which plans and roles apply

SharePoint comes with Microsoft 365 Business, Microsoft 365 Enterprise and Office 365 plans, so there is no separate plan to check. What differs is the role that lets you see each screen.

Why the numbers can differ and what they include

SharePoint counts storage in binary gigabytes, where 1 GB equals 2 to the power of 30 bytes, and the figure includes older file versions and the contents of site recycle bins. It therefore reflects the full history rather than only the current files. Deleted sites that sit in the recycle bin do not count toward the total.

Microsoft notes that the Active sites bar leaves out changes made in the last 24 to 48 hours, and that the storage report refreshes every 48 to 72 hours. The Site usage report can also list different sites from Active sites, because some site templates and addresses are not shown on that page. When the figures disagree, treat the Total usage chart as the authoritative number.

Your quota explains the size you see. Most plans give 1 TB plus 10 GB for each license purchased, and Microsoft 365 F1 and F3 plans give 1 TB in total. That allowance does not include the OneDrive that each licensed person receives, so report OneDrive separately and do not add it here.

03

What should you know about your SharePoint data before you apply?

The storage size

The unit for SharePoint is gigabytes of storage used. Report the figure from the Total usage chart, which already includes file versions and recycle bins, and keep OneDrive on its own line.

The years of history

The years of history run from the year your company started using SharePoint in Microsoft 365, or from the year your oldest team sites were created. If you moved older files over from a shared network drive, count from the start of that history.

The number of seats

Seats are the licensed Microsoft 365 users who have SharePoint access. Your license count also explains your quota, since the allowance grows with each license.

What you don't need to do first

You don't need to export anything to apply, and you don't need to clean or scrub an export. The data vendor anonymizes the records before anyone else sees them, and nothing moves until a lawyer-reviewed license is signed. The application, the intro call and each later step are laid out in how to sell your company's data.

04

How do you export your SharePoint files and history?

Who can run an export

The account doing the export must hold a Microsoft 365 Enterprise E3 or E5 license according to Microsoft, and also a SharePoint E3 license if the SharePoint items are to export properly. An admin grants the permissions by placing the account in the eDiscovery Manager role group, found in the Microsoft Purview portal under Settings > Role groups. Microsoft's guide to exporting eDiscovery search results sets out these rules.

The standard edition of eDiscovery covers sites and files from Office 365 E3 and Microsoft 365 E3 upward, while premium eDiscovery features call for E5 or an add-on license.

The eDiscovery steps

eDiscovery is Microsoft's tool for searching a company's records and collecting them for legal and compliance work. These steps use it to collect whole sites with their full edit history.

1. Place the admin who will export in the eDiscovery Manager role group (Settings > Role groups in the Microsoft Purview portal) and check the licenses described above.
2. Open eDiscovery > Cases, create a case and add a search that points at the SharePoint sites you want. For a large tenant, run Generate statistics beforehand, which lets the export skip sites without matches.
3. Choose Export on the search and pick the export type Export items with items report. If you want the whole edit history, put Select document versions on All versions; the shorter options are Latest version only, Recent 10 versions and Recent 100 versions.
4. Tick Organize data from different locations into separate folders or PSTs. Then tick Include folder and path of the source, so every site keeps its own folders. Where folders are nested deeply, also tick the option that condenses paths.
5. Download the packages from Process manager within 14 days. An export that runs for more than seven days is cancelled, so split big jobs by site or by date range.
6. A small site needs no export: sync its document library with the OneDrive sync app and copy the files out, or pick the files in your browser and choose Download.

Which route suits which size

The three ways of getting files out of SharePoint differ in how much they can carry and how much history they keep.

What a SharePoint export leaves out

An eDiscovery export skips files that a data loss prevention (DLP) rule restricts, which is a company rule that blocks certain sensitive files from being shared. It also skips sites locked with the NoAccess setting and any file the collecting account cannot reach under SharePoint's access controls. Files attached to list items export only if you choose the list expansion option, which Microsoft calls Select items in lists and attachments.

The SharePoint Migration Tool does not help here. It moves content into Microsoft 365 from SharePoint Server and network file shares, and it cannot export content out of SharePoint in Microsoft 365.

What your audit log will show

When eDiscovery collects SharePoint items, Microsoft may record a FileAccessed event in the Microsoft 365 audit log. That event shows that the eDiscovery service opened the files to do the work you asked for, so it is expected rather than a sign of a problem.

How long an export takes

Microsoft gives no fixed time, and exports that include every version take longest. Start early, and plan to split a large tenant into several exports.

Sources: Microsoft Learn, Manage site storage limits in SharePoint, SharePoint storage report, SharePoint site usage report, Export search results in eDiscovery and SharePoint limits. Steps checked against these pages in November 2026.

05

What stays private when SharePoint data is licensed?

You choose what is in scope

An eDiscovery search only collects the sites you put into it. A company can leave its HR, legal and executive sites out and license the rest. The data vendor also deletes records too sensitive to clean safely instead of anonymizing them.

Names are replaced before a lab sees them

Anonymization means replacing every detail that identifies a person, such as a name or an email address, with a stand-in that is the same every time it appears. The document still reads as a whole, so a lab can learn from it, but it no longer says who the people are. The diagram below shows one contract document before and after.

The method, including what is deleted outright and how you review samples, is in how company data is anonymized for AI training.

Ownership and payment

Licensing gives the buyer limited rights to a copy, while the company keeps ownership of the originals. The terms that matter, such as exclusivity, permitted use and deletion, are covered in data licensing. micro1 pays in full within 30 days of accepting the data, usually as a single payment.