Data licensing guide
By system

Licensing email archives (Outlook, Gmail) for AI

By Ade Molajo · Updated · How we check figures

Quick answer

Possibly, but carefully. Mode names email, including Gmail and Outlook, among the records it buys; micro1 doesn't list email. Email is the most personal system a company runs: every message has senders and recipients, and many hold client material or privileged advice. Shared mailboxes and process threads are the realistic scope. Personal, HR and client correspondence stays out.

Mode
names email among its sources; micro1 doesn't
Most personal
system a company runs
Shared mailboxes
are the realistic starting point
Privilege
can be lost if legal advice is disclosed

01Who takes email?

Mode names email among the records it buys, including Gmail, Outlook and Microsoft. It agrees the systems and date ranges with you first, your team signs in to the approved systems, and Mode removes sensitive material and personal identifiers before onward delivery. micro1 lists SOPs, knowledge bases, CRM records and project histories, but not email.

Where these records usually live

Email

  • Gmail
  • Outlook
  • Microsoft 365
  • Google Workspace

Examples of common tools, to help you find your records. Product names and logos belong to their owners; no affiliation or endorsement.

02Why email needs the most care

Every email has a sender and recipients, and most mention other people. Many carry client material, and some carry legal advice, where privilege can be lost if it's disclosed. Staff also use work email for personal matters. That makes email the hardest system to license cleanly, even when a programme de-identifies it.

03What to include, and what to leave out

Email archives: usually licensable (Shared mailboxes for support, ops, projects; Process threads: approvals, handovers; Templates and standard replies; Internal-only threads, people removed) versus check first or leave out (Client email covered by contracts or NDAs; Personal and HR correspondence; Legal advice (privilege); Anything with health or financial account data).Usually licensableShared mailboxes for support, ops, projectsProcess threads: approvals, handoversTemplates and standard repliesInternal-only threads, people removedCheck first, or leave outClient email covered by contracts or NDAsPersonal and HR correspondenceLegal advice (privilege)Anything with health or financial accountdata
A starting point for scoping, not a legal test.

04Scoping email realistically

Scoping Email data in four steps: pick mailboxes (shared ones first), filter out (client, legal, hr), remove people (names, addresses), tell staff (privacy information).01Pick mailboxesShared ones first02Filter outClient, legal, HR03Remove peopleNames, addresses04Tell staffPrivacy information
Four steps before anything is shared. Agree the scope with the programme in writing.
  1. Start with shared mailboxes: support@, ops@, projects@. They're about work, not people.
  2. Exclude individual mailboxes unless you've a clear basis and staff have been told.
  3. Filter out client correspondence covered by contracts or NDAs.
  4. Exclude legal, HR and finance threads by sender, label or folder.
  5. Tell staff. Your privacy information has to say if data is shared or sold. See the ICO's guidance and our UK GDPR guide.

05Simpler alternatives

Much of what's in email is also in tickets, wikis and project records, which carry less personal data. Take our free fit check to see which programmes you're likely to match. The programme decides.

06Questions people ask

Can I sell my company's email archive to AI companies?

Mode buys email as part of an agreed copy. But email is full of personal and client data, so scope it tightly: shared mailboxes and process threads, with people's details removed.

Do we need staff consent?

Not necessarily, but you need a lawful basis and staff must be told. Employee email is personal data under UK GDPR, and staff can object to processing based on legitimate interests.

What about emails with clients?

Leave them out unless your contracts allow it or the client agrees. They're often confidential, and some will be privileged.

Sources

Check which programmes you’re likely to matchA few short questions, about a minute. Free, no sign-up. The programme decides who qualifies.trackimpact.io/data →