Enterprise email archives
Exports of company mailboxes: threads with replies and forwards, attachments, distribution lists and calendar events, from Google Workspace or Microsoft 365. Labs use them to train email and scheduling agents on current business correspondence, which the public corpora from the early 2000s don’t reflect.
What a record contains
One thread: messages with pseudonymized headers, bodies, attachments or their extracted text, and related calendar events.
| Field | What it holds |
|---|---|
thread_id | Pseudonymous ID |
messages[] | From, to, cc (pseudonymous), date, subject, body |
attachments[] | Files or extracted text |
labels | Folders or labels |
related_events[] | Calendar events tied to the thread |
{
"thread_id": "E-77310",
"mailbox_type": "shared",
"messages": [
{"from": "ops_03", "to": ["vendor_x_01"], "date": "2025-01-20T15:02Z",
"subject": "Revised delivery window", "body": "Can we move Thursday's drop to 7am? Dock 2 is closed."}
],
"attachments": [{"type": "pdf", "text_ref": "att_1"}],
"related_events": [{"title": "Delivery", "start": "2025-01-23T07:00Z"}]
}How AI labs use it
- Email agents
- Drafting, triage and follow-up on real business correspondence.
- Scheduling
- Threads paired with the calendar events they produced.
- Document understanding
- Contracts, invoices and decks in the context of the email around them.
Typical preparation requirements
Agreed with the supplier before any work begins. Typical requirements include:
- Personal, HR and legally privileged mail excluded
- External correspondents’ names and addresses pseudonymized or redacted
- Attachments screened for third-party confidential material
Every dataset has a documented owner and confirmed licensing rights. See data governance on sourcex.si.
What makes a strong package
- Shared mailboxes with clear workflows (support, sales, operations)
- Several years of history
- Calendars kept with mail
Compared with public datasets
Public sets such as Enron Email Dataset and Avocado Research Email Collection are useful references, but limited as enterprise training data. The chat and email category page compares them with licensed data.
Who typically holds it
- Professional services firms
- Logistics companies
- Real estate firms
- Construction companies
Know a company like this?
Introduce the company to SourceX. If its data deal closes, you can earn up to $100,000 in referral fees, paid after the buyer accepts the data and SourceX receives payment.
Refer a companyQuestions
Whose mailboxes are included?
That’s agreed in scope. Shared and role mailboxes are the usual starting point; individual mailboxes only where policies and notices allow.
How are people outside the company handled?
Their names and addresses are pseudonymized or redacted before delivery.
Need this data for a model?
Describe what you need: domain, volume, history, format and licensing terms. SourceX looks for companies that hold it and can license it.