Can a company license its chat and email archives for AI training?
Yes, if your company owns the archives, works within its platforms' terms and protects the people in them. Channels and email threads are where a company's decisions actually get made: a request arrives, people discuss it, someone hands it off, a document is attached, and the thread records who decided what and why. Much of a company's data already lives in third-party software. Licensing it means re-licensing what already exists, and getting paid for it rather than giving it away for free. For conversations, that software is Slack, Microsoft Teams, Google Chat, Gmail or Outlook, and the archive goes back as far as your retention settings have kept it.
Why do AI labs and agent builders value workplace conversations?
Because agents that work inside companies have to learn how work moves between people, and that is recorded almost nowhere in public. The Enron corpus, a public set of corporate email, holds about 0.5 million messages from about 150 users and was made public by the Federal Energy Regulatory Commission during its investigation. Its distributor describes it as, to their knowledge, the only substantial collection of real email that is public, because privacy concerns keep the others private (Enron Email Dataset, Carnegie Mellon University).
Buyers want the sequence, not single messages. Business Insider quoted an executive at an AI training-data startup saying AI developers increasingly want "long-horizon" data that gives a full picture of how work gets done; and that if a company will not sell its Slack, CRM or email records, the rest of its corpus is less valuable (Business Insider, Oct 13 2026). What conversations teach:
- Decisions. Threads where options were weighed and someone chose, with the reasons written down.
- Handoffs. A request passed from sales to engineering to support, with the context that travelled with it.
- Procedures in practice. The "how do I…?" question in a help channel, and the answer that became the standard way of doing it.
- Edge cases. Incidents, escalations and the exceptions a manager approved.
What makes chat and email archives more valuable?
Archives are worth more when they cover a long period, keep their structure and capture real variety. In rough order of impact:
- Variety and depth. The person who leads agent training environments at an AI training-data company told MIT Technology Review that one company's Slack history, if it captures conversations with variety and depth, can be worth more than another's Slack, email and calendars combined (MIT Technology Review, Oct 13 2026).
- Span. Years of continuous history rather than a recent slice.
- Structure intact. Threads, replies, timestamps, channel names, roles and links to attachments, as the platform's export keeps them.
- Linked systems. Messages that point to tickets, documents, pull requests or CRM records that can be joined to them.
- Breadth. The chief executive of a startup that sells data to AI developers told Business Insider that the more white-collar employees a company has, the more valuable its data corpus is (Business Insider).
What gets removed or de-identified from chat and email?
People are in every message, so exclusions and de-identification matter more here than for most kinds of data. You approve the scope, and you decide which channels, mailboxes, teams or periods to propose leaving out of it. fearofgod.ai de-identifies the data before it goes to buyers. Raw data is deleted according to a schedule agreed with the partner.
Typical exclusions you might decide on:
- Sensitive spaces. Direct messages, HR, legal and board channels, and the mailboxes of people who handle sensitive matters.
- Privileged and client material. Advice from lawyers, and clients' confidential information covered by an NDA or engagement letter.
- Personal matters. Health, family and other private messages, and anything about people outside the company.
- Secrets. Passwords and keys pasted into chat.
Pseudonyms alone do not take messages outside European data-protection law. The GDPR treats pseudonymised data that could be linked back to a person with additional information as information about an identifiable person; only data rendered anonymous falls outside it (GDPR, Recital 26). This is general information, not legal advice.
Your company's name stays confidential until the agreement is fully signed, and a completely anonymous partnership is also possible. Your company can keep ownership of its data. This is agreed case by case.
What rights and consent do chat and email archives need?
The gates come from three directions: the platforms, the people and your clients. This is general information, not legal advice.
- Slack's API terms. An app offered for use outside your organisation may not use API data, meaning messages, files and metadata from Slack's APIs, to train a large language model, and may not bulk export Slack message and file data except where an additional agreement expressly allows it (Slack API Terms of Service, effective Oct 10 2025).
- Slack's own export. Free and Pro workspaces can export public channels but not private channels or DMs. On Business+, owners apply for a self-serve tool that exports private channels and DMs "for legal and regulatory compliance", and Enterprise org owners request export tools from Slack's support team (Slack help). Read your own Slack agreement before exporting anything for a licence.
- Employees. Whether consent given at work extends to a sale of employees' messages is, in one law professor's words to Business Insider, "an ethically gray area" (Business Insider). Lawmakers are watching: on Oct 8 more than 120 members of Congress wrote to the parties in a bankrupt airline's proposed data sale, urging them to protect former employees' privacy (Rep. Steven Horsford, Oct 8 2026).
- Clients and counterparties. Email with clients, suppliers and lawyers carries their confidential information, and NDAs and engagement terms can rule whole mailboxes out.
What are chat and email archives worth?
Published figures span orders of magnitude, because they mix a bankruptcy auction, a small wind-down sale and a platform licensing years of public posts. Each is dated and labelled by market side; none is a fearofgod.ai price.
| Reference point | Market side | Figure and source |
|---|---|---|
| Bankrupt airline's emails, Teams messages, code and records | Court-process bids, Sep 2026 | $10 million bid, then a $12.5 million proposal; another bidder offered $7.5 million; no sale court-approved as of Sep 14 (Fortune) |
| Same archive, for scale | Reported size, Oct 2026 | about 100 million emails and 500 million Microsoft Teams messages, per court findings cited in a congressional letter (Rep. Horsford) |
| A wound-down startup's Slack and workspace data | Reported sale, Oct 2026 | about $15,000 pending for workspace data in Google, Notion, Figma and Slack, after roughly $25,000 for the code (MIT Technology Review) |
| A forum platform's conversation data | Company filing, Feb 2024 | licensing arrangements signed in Jan 2024 with an aggregate contract value of $203.0 million and terms of two to three years (Reddit S-1) |
Business Insider also reported owners receiving six-figure "exploding offers" that expire in 48 hours (Business Insider). A deadline that short leaves no time to check platform terms, employees or clients. Partnerships typically range from $50K to $3M+.
How does licensing chat and email archives work with fearofgod.ai?
You license conversations your company already keeps, securely, and earn additional, long-term revenue from them. They already sit in your chat and email systems, so there is nothing new to collect. Your company can be paid more than once, through renewals and repeat licences.
The steps are the partnership form, a conversation, scope and terms, data preparation, then delivery and payment. A useful first description covers the platforms, the years of history, a rough headcount, which channels or mailboxes you would consider, and whether threads link to documents or tickets. Talk to our team when you are ready. Conversations often sit next to other records worth licensing: see licensing support tickets and licensing codebases.
Selling chat and email archives to AI companies
A company that owns its workplace conversations can sell its chat and email archives to AI companies by licensing them with fearofgod.ai, securely and for long-term revenue. Partnerships typically range from $50K to $3M+. Partners can be paid more than once, through renewals and repeat licences.