Selling a company's email archive for AI training: exports, consent and price

Yes: a company can license its own email to an AI buyer, and email is one of the systems buyers ask for by name. A buyer's revenue chief said in October 2026 that if a company will not sell its Slack, CRM or email records, the rest of its corpus is worth less. Both big mail suites let the company take its own mail out: Google Workspace offers a full organisation export in all editions to a super administrator, and Microsoft's data protection terms say the customer retains all right, title and interest in its data and may extract it during the subscription. What both forbid is a third-party app pulling the mail through their APIs to train AI, so the export has to be your own. The obstacles are inside the mailboxes: employees, customers, other companies' confidential material and privileged advice. Buyers price mail as part of the company's whole record: the median of the per-company sales on record is $28.3K per company, and one broker's illustration values a logistics company's 2.1 million Gmail messages at $41,000 within a larger package. The steps are: export it yourself before any subscription lapses, take out what cannot be licensed, de-identify the rest, and put a description of it before every buyer that fits at once.

This guide is for a founder, officer or trustee of a company, operating or closed, whose mail sits in Google Workspace or Microsoft 365. It covers what a company's mail holds and what buyers value and leave out, how each suite lets an administrator export the whole organisation's mail and what the records say about the deadlines when a licence lapses, the law on employees' and customers' messages, privilege, confidentiality owed to other companies and earlier privacy promises, de-identification of names, addresses and signatures, what mail fetches as part of a company's records, preparing and delivering, the case of a closed company, and how a sale through AIData.Broker works. Every quote comes from the page or filing it names, read on the day shown in its record.

This guide is research from public sources, not legal advice.

Contents

1. What a company's mail holds, and what buyers value in it
2. What buyers leave out of a mail export
3. Exporting a whole organisation's mail from Google Workspace
4. Exporting a whole organisation's mail from Microsoft 365
5. The deadlines when a subscription lapses
6. The law on employees' messages
7. Customers, other companies, privilege and earlier promises
8. De-identifying mail: names, addresses and signatures
9. What mail fetches as part of a company's records
10. Preparing and delivering a mail export
11. A closed company's mail
12. How a sale through AIData.Broker works
13. Questions people ask
14. What this means for you

1. What a company's mail holds, and what buyers value in it

Email is the oldest and widest record a company keeps. Chat holds the inside of the company: quick questions between colleagues. Email holds the edges and the formal record: the proposal sent to a customer and the customer's reply, the negotiation with a supplier, the board update, the hiring loop, the incident report, the decision written up so that it could be forwarded. A company that has run for ten years has ten years of that, from every role, in one system, and it is often the only system that reaches back to the company's first day. That breadth is what buyers want. Models that work inside companies need to learn how real work moves: who asks whom, how a request turns into a decision, how a problem raised by a customer becomes a fix and a reply. Public text on the web has almost none of it, because companies do not publish their mail.

Buyers say plainly that communications sit near the top of what they want. Turing, which buys companies' whole working records, answers the question of whether it only wants code by saying that communications, tickets, wikis and databases stand on their own and are often the most valuable part. Polyshares, which buys companies' operating records for its own account, names Slack, email, tickets, CRM and files as the record it prices. Corpus, a broker, prices a company's workspace data: email, chat, meetings and CRM. Mode, through its BizBuyData service, describes the operating history a business already owns: email, documents, chat, calls, CRM, billing and operations. And the largest records sale on record, Spirit Airlines' sale to Google, covered about 100 million emails across 80,000 email accounts, with Teams, OneDrive and SharePoint beside them.

Four things in a mail archive move its value most.

Threads that end in an outcome. A reply chain that starts with a request and ends with an answer, an agreement, a shipped change or a resolved complaint is a complete example of work. A newsletter, a notification from a tool or an unanswered announcement is close to worthless on its own. Corpus Bridge, a broker for communications data, counts outcomes attached and clean threading, with each reply attached to the right question in the right order as adding value, and languages other than English as well. Mail threads are messier than chat threads: replies quote the whole history below them, people fork a conversation by changing the recipients, and forwarded chains arrive inside new ones. An export that keeps the message headers intact (who replied to which message, and when) keeps that structure recoverable.

Decisions written down. Email is where companies record what they decided, for people who were not in the room: the summary after a meeting, the approval of a budget, the reasons a product was cut. These messages are short and dense, and they show the reasoning of experienced people in their own words. Buyers pay more for the work of white-collar specialists than for general traffic: Mode.Inc's chief executive told reporters that value rises with the number of white-collar employees.

Correspondence with customers and suppliers. The threads with people outside the company are where its real work shows most clearly: a quote, a dispute, a renewal, a technical question answered over several replies. They are also the threads with the most rights attached, because the other side never agreed to anything (section 7). They are the most valuable and the most constrained part of a mailbox at once, which is why the way you handle them decides much of the price.

Years and links. Estimators price years directly: Ooak's estimator values a 51 to 100 person operating company with four or more years and six tools, Gmail among them, at $85k to $121k, and the same company with seven or more years at $99k to $139k. Links matter as much: a message that links a ticket, a document or a pull request ties the conversation to the work product, and a wind-down firm told Forbes that a record tied to the work it produced is worth more than a standalone document. The link only carries value if the other end comes too, which is why buyers price mail beside the documents and tickets it points to. A buyer's own lever list says it bluntly: bring every connected system, especially documents and email.

What lowers value is the opposite: a domain that mostly received automated mail, a few years of history, mailboxes of people who left and were deleted, or a company whose real work happened elsewhere. A plain count before you talk to anyone, split by mailbox, year and the share of mail sent to or from outside domains, tells you most of what a buyer will ask first. Polyshares' system guide shows where to find the size for Microsoft: the Outlook mailbox usage report in the admin centre.

2. What buyers leave out of a mail export

A mail export holds everything at once: the work, and the material nobody may license. Buyers and their counsel expect the second kind to be found and excluded before anything leaves, and one law firm advising holders on these sales tells them to exclude high-risk data entirely rather than relying on de-identification. The categories that come out of a mail archive are the same in almost every company.

The list in the buyers' own words is long. The rights question on counterparty confidentiality quotes a law firm's checklist: privileged material, HR and personnel files, payroll and tax records, health and benefits information, union-related communications, customer and partner confidential information, licensed data, and data from regulated systems. For mail, most of that list can be found by who the message was with (a counsel's domain, the HR mailbox, a supplier's domain), by which mailbox it sits in, and by a content screen for the rest (section 8).

Exclusion does not have to cost much value. The Spirit sale is the reference case: after the auction, the estate and Google agreed that passenger transaction, refund and inflight sales systems would not be transferred in any form, which left corporate email in scope under an independent de-identifier, and the sale is still described at the same price. Excluding what cannot be licensed is part of what makes the rest sellable.

3. Exporting a whole organisation's mail from Google Workspace

Google controls the gate for Gmail, and its admin help page sets out how the whole organisation's data leaves. The full reading is in our Google Workspace terms record.

Who may export. Google says the full export of an organisation's data is available in all Google Workspace editions, and that it needs a super administrator account that is at least 30 days old, with 2-Step Verification on. Two consequences follow. The export is run from the company's own admin console, by the company's own top administrator, which is exactly the route buyers' counsel want to see ( the vendor's native export tool, run by your own admin). And the account matters: if the only super administrator has left, or the account was created last week to run the export, you wait. Make sure a current officer holds a super administrator account well before you need it.

How long it takes. Google says the export is available no earlier than 48 hours after you start it. For a large domain it can take longer. Start it as soon as you decide to sell, not when a buyer asks.

Large and regulated organisations, and narrower exports. Google's record says organisations with more than 1,000 users, or with FedRAMP authorisation, must contact support first. An export of specific users, or a filtered export, is offered "if available", and the advanced editions such as Enterprise Plus and Education Plus use a separate flow to choose which Workspace data to export. Google Vault, on the editions that include it, is the eDiscovery export route, and it is often the more practical tool when you want only some mailboxes or a date range.

What the API terms forbid. Google's API User Data and Developer Policy, last updated 3 September 2026, prohibits an app that obtains Workspace data through Google's APIs from "transferring, selling, or using user data to create, train, or improve a machine learning or artificial intelligence model beyond that specific user's personalized model". The same policy forbids transferring or selling user data to data brokers or any information resellers. That clause binds apps: a buyer's connector, a broker's sync tool, any service that asks your administrator to authorise it on your domain. It does not govern your own admin export. Some buyers offer read-only connectors to source systems ( Labelbox's Prism says it uses them); for Gmail, the safer course is the export your administrator runs, with the files then licensed under your own agreement. Of the 18 business systems whose terms we have read, 6 now bar training AI on what an app pulls through their API, and Google Workspace is one of them.

What Google's records do not say. Our record does not hold the file format Google uses for exported mail, how deleted users' mail is treated in the export, or how long Google keeps a removed user's mailbox. Read the current admin help page for those before you plan, and export the mailboxes of departed staff while they still exist.

4. Exporting a whole organisation's mail from Microsoft 365

Microsoft's position starts from ownership. Its Products and Services Data Protection Addendum, dated May 2026, says that "as between the parties, Customer retains all right, title and interest in and to Customer Data", and that at all times during the subscription the customer "will have the ability to access, extract and delete Customer Data" stored in each online service. The full reading is in our Microsoft 365 terms record.

The route for a whole organisation. Bulk export of mailboxes and Teams chats for a whole organisation is usually done through Microsoft Purview eDiscovery. Polyshares' system guide names the same route for Outlook: a Purview eDiscovery export to PST. The licence matters: Microsoft Learn says premium eDiscovery features require Office 365 E5 or Microsoft 365 E5, or the E5 add-ons. A vendor's guide says standard content search and export are included from E3, but our record notes that this was not verified against Microsoft's licensing documents, so confirm what your own tenant includes before you plan around it.

Who runs it. The export is a compliance tool inside your own tenant, run by your own administrator with the eDiscovery role, which again is the route buyers' counsel want ( the export route). Turing's intake accepts records only from someone with admin authority, so the person who can run the export is also the person who should start the conversation with any buyer.

What the API terms forbid. Microsoft's API terms, last updated October 2025, cover Graph, Exchange, Teams, SharePoint and OneDrive. An app may not "scrape, build databases or otherwise create copies of any data accessed or obtained using the Microsoft APIs" beyond what its intended scenario needs, and may not redistribute, resell or sublicense data obtained through them. The record adds that email protocols and APIs may be used only for syncing or backing up mail, calendar and contacts unless customers expressly grant other permissions. A change to the Product Terms on 1 May 2026 says that outputs generated through robotic process automation, bots or similar technologies may not be used to train AI models; it is aimed at automated outputs of the Office apps and does not, on its face, bar a customer from licensing its own mail. The route that works is the one that keeps every connector out: your own eDiscovery export.

What Microsoft's records do not say. Our record does not hold how Microsoft treats the mailbox of a user whose licence is removed, how inactive mailboxes or litigation holds affect an export, or the size limits of a single export job. Check those against Microsoft's current documentation, and place a hold or export the mailboxes of departing staff before their licences are removed.

5. The deadlines when a subscription lapses

The most common way a company loses its mail is by stopping payment. A closing company cancels its subscriptions to save money, and the vendors delete on their own schedules. The records say this about each suite.

Sources: Google Workspace admin help on cancelling, and Microsoft's data protection addendum and Microsoft Learn on retention, each held in the Google Workspace and Microsoft 365 records. Our deletion clock counts the days left for each system from the date the subscription ends.

Two points follow from that table. For Google, there is no stated grace period in which a cancelled domain can be exported: the export has to happen before the cancellation. For Microsoft, the window is longer, but the useful part of it is the time administrators can still reach the tenant; do not plan to export in its last weeks. In both cases the export itself takes time (Google's takes at least 48 hours), and a large archive takes longer to download and check than to export. A company that may sell its records should keep the mail subscription paid until the export is complete and verified, and treat that cost as part of the sale.

Records also lose value with time in a softer sense. Turing's intake warns that records depreciate fast after operations cease, naming Slack and CRM; its record does not name a figure for email, but the logic is the same: the people who could answer a buyer's questions leave, the linked systems are shut down, and the context goes with them.

6. The law on employees' messages

Ownership is rarely what stops a mail sale. Under US copyright law, a work " prepared by an employee within the scope of his or her employment" is a work made for hire, and the employer is considered the author. Microsoft says the same from its side ( the customer retains all right, title and interest). What stops a sale is the people in the messages and the promises made to them. The employee messages guide treats this in depth; what follows is what matters most for email.

Federal law. The Stored Communications Act governs messages at rest. It is an offense to access stored communications without authorization, or by exceeding an authorization, but the access offense does not apply to conduct authorized "by the person or entity providing a wire or electronic communications service". A company that provides email to its own staff can authorize access to what is stored on that system, and where Google or Microsoft hosts it, the company exports under its contract through the vendor's own admin tools. The Act's ban on divulging stored contents applies only to a provider of service "to the public", which a company's own mail system is not; it does bind the hosting vendor, which may disclose contents only within exceptions such as the lawful consent of the originator or an addressee, one reason vendors release mail only to the customer's administrators. The Wiretap Act governs mail captured in transit, and it makes it an offense to use the contents of a message knowing or having reason to know it was unlawfully intercepted. Licensing an archive that already sits in the mail system is a stored-communications question, not an interception one; letting a buyer install a tool that reads live mail would turn it into one.

Two exceptions are often read too broadly. The provider exception lets a provider's staff intercept or use communications in the normal course of employment when that is a necessary incident to providing the service; licensing staff mail for revenue is not. The one-party consent rule covers interception where one of the parties has given prior consent; an employee's consent speaks only for that employee, and the customer on the other end has not consented.

Monitoring notices. Several states require notice before an employer monitors employees' electronic communications: New York's Civil Rights Law 52-c, Connecticut's 31-48d, Delaware's 19 Del. C. 705 and Maine's 26 M.R.S. 620-A. They govern notice, not licensing, but pulling every mailbox into a review is close to monitoring, and a company that never gave a required notice has a weaker answer to every later question about its staff's expectations. Read your acceptable-use policy for what it actually told staff: a notice written for security reviews is evidence of the company's authority over its systems, not consent to a licence.

California. California staff have had full privacy rights over their work data since the state's employee exemption became inoperative on 1 January 2023. The state's privacy agency says consumers include employees, job applicants and contacts at business customers and vendors. Licensing personal information for money is a sale under the Act's definition, "for monetary or other valuable consideration", and the Act treats the contents of a consumer's mail, email and text messages, unless the business is the intended recipient as sensitive personal information. That clause was written with email in mind, and whether an employer counts as the intended recipient of mail its staff exchange with outsiders is not settled. Buyers' counsel ask how California employees' rights were handled; most holders answer by removing employee personal information to the Act's deidentified standard, with a public commitment and contracts binding every recipient to the same.

The EU, the UK and Canada. Staff in the EU and UK need a lawful basis for the new use, and consent from employees is rarely valid given the dependency of the employment relationship. For customers or staff in Canada, the question buyers ask is whether each person consented to the new use or the data is anonymised to the Canadian standard before it leaves. A holder with staff abroad should expect those mailboxes to need fuller anonymisation than US ones.

The lesson from Spirit. Employees are what slowed the largest records sale on record. Spirit's archive included about 100 million emails, and 119 members of Congress asked Spirit and Google to halt the transfer of former employees' data until protections were in place. The court's privacy ombudsman wrote that she did not investigate or reach any conclusion about risks to employees, and the sale hearing was adjourned five times. A holder outside bankruptcy has no ombudsman and no court to settle these questions, so it settles them before the licence.

7. Customers, other companies, privilege and earlier promises

Employees are the people who wrote most of the mail. The people who received it, and the promises the company made to them, decide the rest of the scope.

Customers. Customer correspondence is some of the most useful mail a company holds and the most constrained. If your terms of service or data processing agreement say you process customer content only to serve the customer, licensing that content to an AI buyer breaks the promise, and under European law a processor that decides the purposes of processing for itself is considered a controller in respect of that processing. Buyers' counsel ask that customer content, including customer email, be excluded, or that each customer contract expressly permit the use, or that the data be anonymised to the standard and the contract not forbid it. For a business that sells to consumers, the question is also whether the privacy policy in force when the mail was collected allowed this kind of transfer. The FTC's technology blog, quoted in that question, put it this way: a business that collected data under one set of privacy commitments cannot then unilaterally renege on them, and a retroactive change to a policy may itself be unfair or deceptive. A new policy today does not reach mail collected under the old one.

Other companies. A supplier's pricing, a partner's roadmap, a customer's draft contract: mail carries other companies' confidential material constantly, often received under an NDA that limits its use to the relationship. Buyers' counsel ask that external email threads, partner material and attachments from third parties be excluded or reviewed, and that you can list which counterparties appear. The Spirit sale drew objections from vendors who said their own data inside Spirit's email, Teams and SharePoint was not estate property, with motions to enforce their confidentiality agreements. Email makes this list easier to build than any other system, because every message names the domains it went to and came from. Count messages by outside domain, mark the counterparties you hold NDAs with, and exclude or review those threads.

Privilege. Communications with your lawyers for legal advice are protected only while they stay confidential, and a voluntary transfer to a licensee is the kind of disclosure that waives protection. The federal rule that rescues mistakes is narrow: it protects an inadvertent disclosure made in a federal proceeding or to a federal office or agency, not a commercial licence. Buyers ask that outside counsel email and litigation folders be excluded. In a mail archive, privilege is usually found by the lawyers' domains and addresses, by an in-house counsel's mailbox, by "privileged" markers in subject lines, and by a reviewer's pass over what those filters miss. Your own trade secrets are a business decision rather than a legal bar: decide what you are willing to have a model learn, exclude the rest, and restrict the buyer's use in the licence where you keep material in.

Licensed content and vendors' own training. Mail carries attachments the company licensed from others (research reports, purchased lists, stock images), which buyers ask to be reviewed against the licences. Buyers also ask whether any vendor has already trained its own AI on the content, because exclusivity is worth less if it has. Our records do not hold Google's or Microsoft's own statements about training on customer mail; check each suite's current AI settings and note them in your rights file.

The rights file. Each of these questions becomes one page of a rights file: who signed, how the mail was exported, which domains and mailboxes were excluded and why, what the policies in force said, and who de-identified it. Buyers' counsel read it before they price, and a holder that has one sells faster.

8. De-identifying mail: names, addresses and signatures

Every buyer that takes company communications expects personal data to be removed before or as part of delivery. The question is who does it, to what standard, and when. Buyers differ. Scale says it anonymises inside your own systems before anything is extracted. Turing removes personal data and secrets itself, in isolated processing, and you approve the scope first. Polyshares says that identifying fields, customer names, emails and account numbers are stripped by a specialist partner before it or any buyer sees the record. Corpus Bridge says the standard is negotiated with the buyer and you approve it before delivery. In the Spirit sale, the estate and Google voluntarily agreed to follow the deidentification standard adopted in the California Consumer Privacy Act, and Google's opening bid already provided for a third-party deidentification process for which the buyer would bear the cost. Each model except the first means someone other than you reads raw mail; decide which you accept before you choose a buyer.

Email has shapes that a de-identifier must handle, and most of them are not in the body of the message.

• Headers. Every message carries the sender, the recipients, the copied and blind-copied addresses, display names, and often the names of the mail servers it passed through. Replace each person with a stable pseudonym, consistent across the archive, so a buyer can still see that the same person replied to the same colleague many times, which is what makes threads valuable, while the names and addresses are gone.
• Addresses everywhere. Email addresses appear in headers, in the body ("cc'ing jane@"), in quoted replies, in forwarded chains, in calendar invitations and in links. Phone numbers and postal addresses appear in signatures and in the text.
• Signatures. A signature block is a compact identity card: name, title, company, direct line, mobile, office address, sometimes a photo or a pronoun line. It repeats in every message a person sends, and in every reply that quotes them. Remove signatures as a block rather than token by token, and remove the legal disclaimers below them, which add bulk and often name the company.
• Quoted history. A reply quotes the message before it, which quotes the one before that. The same personal details recur many times in one thread. Either de-identify every copy consistently or strip quoted history and rebuild the thread from the headers.
• Attachments. Contracts, CVs, invoices and spreadsheets carry more personal and confidential information than the messages that send them. Decide whether attachments are in scope at all; if they are, they need the same treatment as documents.
• Secrets. Staff send passwords, keys and account numbers by email. They come out whatever the buyer's standard.
• Sensitive topics. Health, family, beliefs and pay appear in ordinary mail. A content screen flags records about health, children, beliefs, union membership, finances, HR matters and legal advice for a person to read and leave out, rather than masking them line by line.

Our de-identification standard sets out the method we advise: removal before anything leaves you, by you or an independent provider you choose, with a group-size floor of 11 for any delivery and 300 random records read by a person before it goes. Our free de-identification tool runs in your own browser, so the mail never leaves your computer, and on company Slack, email and documents it measured an F1 of 86.8%, finding 95.8% of people's identifiers. No tool is perfect, which is why the standard pairs it with a person's review.

De-identification has limits that matter more for mail than for any other record. It does not cure privilege (the waiver concerns the content, not the names), it does not cure confidentiality owed to another company (its pricing is still its pricing without a name on it), and it does not make a customer contract permit a use it forbids. So the order of work is exclusion first, then de-identification, then a person's review of a random sample.

9. What mail fetches as part of a company's records

No buyer we have read publishes a price per email or per mailbox alone. They price the company's whole working record, in which mail is one of the systems, and they price it higher when mail is included. Here is what the public record shows for records that name email.

Across every public offer for one company's records, the middle half sits at $174K to $693K per company. The middle of the buyers' offers is $316K per company, and the median of the sales that actually happened is $28.3K per company. Published ranges sit above what closes, partly because sales are disclosed more often for small closed companies; the prices guide explains the gap.

Three patterns in those rows matter for a holder of mail.

Mail is a large share of a package, not a package on its own. In Corpus Bridge's illustration, the mail is the largest single line, ahead of Slack and the ticket system, but it is still about a quarter of the whole. Ooak's estimator shows the same effect from the other side: six tools including Gmail and Drive at $85k to $121k, against $16k to $22k for the same company with only Jira, GitHub and HubSpot. Communications are what make the rest of a company's records worth more, and buyers price them together.

Closed companies get less. The market pays for a closed or closing company's records about 0.65 times what it pays for the same company operating, in part because closed companies sell in a hurry and their records are often incomplete. For mail the main cause is the deadlines in section 5: mailboxes deleted with departing staff, or a domain cancelled before anyone exported it.

The first offer is rarely the best. Business Insider reported in October 2026 that small businesses receive six-figure offers that expire in 48 hours for their internal records, email included. Buyers quote very differently because they buy on different theses, which is why refusing a short deadline and collecting several offers is the single most useful thing a holder can do. Our valuation puts a range on your own company from these records before you talk to any buyer, with no upload.

10. Preparing and delivering a mail export

The preparation guide sets out the full order for any company's records. For mail, it comes down to these steps, most of which need no buyer at all.

1. Secure the administrator and the subscription. Confirm that a current officer holds a Google super administrator account at least 30 days old, or a Microsoft administrator with the eDiscovery role, and that the subscription stays paid until the export is complete.
2. Count before you export. Number of mailboxes (current and former staff), messages per year, total size, the share sent to or from outside domains, and the main outside domains. Polyshares and Turing both start from counts per system, read from the admin console, and Turing says estimates are fine at that stage. Turing also accepts older systems, naming Lotus Notes among them, so a company with mail from before its current suite should count that too.
3. Set the scope by mailbox and by domain. Mark mailboxes in or out (the HR mailbox, in-house counsel and executives' personal assistants usually come out), and mark outside domains in or out (law firms, customers whose contracts forbid the use, counterparties under NDA). Decide whether attachments and calendars are in scope.
4. Run the export yourself. Google's full organisation export or Vault; Microsoft's Purview eDiscovery. Run by your own administrator, from your own tenant, with no third-party app authorised to read mail. Record the date, the tool, the person and the settings.
5. Check the export is whole. Compare message counts in the export with the counts from step 2, mailbox by mailbox. Keep the headers intact: they carry the thread structure buyers pay for.
6. Exclude, then de-identify. Remove out-of-scope mailboxes, domains and threads from a working copy, then de-identify the rest to the standard you chose (section 8). Draw any sample for a buyer from the de-identified copy, never from the raw export, and never show a sample before the rights review, since a sample is itself a disclosure.
7. Write down how you did it. Who exported what, when, which mailboxes and domains were excluded and why, the policies in force, who de-identified and how it was checked. This is your rights file and your proof of title: buyers have been offered archives by people who did not have the rights to sell this data.
8. Agree delivery and acceptance in the licence. The format, who de-identifies and to which standard, objective acceptance criteria and a review window. Never send mail or samples by email or chat: DataVendor tells holders plainly not to send data or code by chat or email, and a mail archive moved that way is a breach waiting to happen.

Keep two things apart: what you describe to buyers, and what you deliver. Buyers price from the description first: counts, years, mailboxes, the share of mail with outside parties, what links to what, and what stays out. Polyshares says it reads an anonymised sample before it prices. The full export moves only after a licence is signed, and only the scope the licence names.

11. A closed company's mail

A closed company is a common holder of a valuable mail archive, and the one with the least time. Three questions decide whether its mail can still be sold: does the mail still exist, who can sign, and who can run the export.

Does the mail still exist? For Google, a cancelled domain's data will be deleted and cannot be restored, with no stated window in which it can still be exported. For Microsoft, the customer can extract data for 90 days after expiration or termination from a limited function account, and the data is gone within 180 days at most. So the first act of a wind-down that may sell its records is to keep the mail subscription paid until the export is done, or to run the export before cancelling. The deletion clock shows the dates for each of your systems.

Who can sign? The company sells, not its founders personally, and the right person signs for it. Buyers' counsel ask for the company that collected the data to be the licensor, with authority approved, and, for a wind-down, the board's approval and any investor consents. In Delaware, where many US startups are incorporated, a dissolved corporation continues for three years to wind up its affairs, including disposing of its property, and it adopts a plan of distribution that provides for its claims; a sale of data by a dissolving company is an ordinary sale under its own privacy policy and the law. In a chapter 7 bankruptcy the trustee signs, because the trustee's duty is to collect and reduce to money the property of the estate, and a sale that includes personally identifiable information runs through section 363(b)(1) and its privacy conditions. Buyers ask that the right officeholder signs. Our estates page sets out the path for trustees, assignees and dissolved companies, and the guide to SimpleClosure and its alternatives compares the wind-down services that also sell records.

Who can run the export? The super administrator or the Microsoft global administrator may have left. Transfer the role to an officer or the trustee while the subscription is live, because a new Google super administrator account has to be at least 30 days old before it can run the full export. If nobody with access remains, the vendor's support is the route, and the company will need to show it still exists and who acts for it.

A closed company's mail carries the same people questions as an operating one, with one difference: the staff are gone and harder to reach. That pushes most closed companies toward thorough de-identification and generous exclusion rather than consent. The value is still real: the cielo24 records, built over more than a decade of Slack, Jira, email and Drive, sold for what SimpleClosure describes as mid-six figures during the wind-down.

12. How a sale through AIData.Broker works

AIData.Broker represents companies selling their data to AI buyers. We act only for the holder, as its agent in the sale, for holders based in the United States, the United Kingdom and Canada. For a holder of company mail, a sale through us runs like this.

The valuation comes first, with no upload. The valuation asks about your systems (mail suite, years, mailboxes, which domains and mailboxes are in scope, what else the company ran) and returns a range from the public records in section 9. If you add files, they are measured on your own computer. The valuation binds no one.

You keep the mail and deliver it yourself. We do not hold your export, show any part of it to a buyer, or license it in our name. You, or a de-identification provider you choose, deliver directly to the buyer, and the buyer pays you directly. That keeps the export on the route both suites' terms allow: your own administrator's export, never a connector.

Every buyer that fits sees it at once. Every buyer whose published requirements fit your records receives a description that does not identify you; each that gives its confidentiality undertaking receives the same datasheet, with one deadline for offers. Offers come back on the same terms, so you can compare price, de-identification model, payment timing and exclusivity side by side. Our free offer check reads any offer you already have against the published terms of the buyers we track.

The fee is due only out of money you receive. It is a success fee on what you are actually paid, band by band: 15% of the first $250,000, 10% from $250,000 to $1 million, and 6% above $1 million. On a $300,000 sale the fee is $42,500. Mandates are exclusive for 2 months, with a tail of 12 months for buyers we introduced, and we do not take mandates where the expected sale is below $26,000. Several buyers publish an introduction fee; in a sale we run we decline it and ask the buyer to add it to your price instead, so we earn the same whichever buyer you choose. How it works sets out every step and term.

Questions people ask

Can a company sell its email archive for AI training? Yes, if the company exports its own mail through its suite's admin tools, takes out what cannot be licensed (privileged, HR, customer and counterparty material), de-identifies the rest, and the right person signs for the company. Buyers name email among the systems they want ( Polyshares, Corpus), and one buyer's revenue chief said that without email the rest of a corpus is less valuable.

Do Google's and Microsoft's terms allow it? Both let the company export its own data: Google's full organisation export is available in all editions, and Microsoft's terms say the customer retains all right, title and interest and may extract its data. Both restrict apps: Google's API policy bars apps from using Workspace data to train AI models or transferring it to data brokers, and Microsoft's API terms bar apps from copying or reselling data obtained through its APIs. We have found no clause in either suite's records that bars a customer from licensing mail it exported itself.

How do we export every mailbox in Google Workspace? A super administrator whose account is at least 30 days old, with 2-Step Verification, runs the full organisation export from the admin console; it is ready no earlier than 48 hours after it starts. Organisations over 1,000 users contact Google support first, and Vault is the route for narrower exports.

How do we export every mailbox in Microsoft 365? Usually through Microsoft Purview eDiscovery, which exports to PST ( as Polyshares' guide describes). Premium eDiscovery features need an E5 licence or add-on; confirm what your own tenant includes.

What happens to our mail if we stop paying? Google deletes a cancelled domain's data with no stated recovery window. Microsoft keeps it in a limited function account for 90 days so the customer may extract it and deletes it within 180 days at most. Export before the subscription ends.

Do employees have to consent? US law does not generally require it for mail on a system the company provides, because the Stored Communications Act lets the provider authorize access ( section 2701(c)). But notice laws apply in several states, California staff have privacy rights over their work data ( California employees), and EU and UK staff need a lawful basis. Read the employee messages guide.

Can we include mail with customers? Only if your customer contracts permit it or the mail is anonymised and the contracts do not forbid that; buyers' counsel ask that customer email be excluded otherwise, and that the privacy policy in force at the time allowed the transfer.

How much is a company's email worth? Buyers price the whole working record rather than mail alone. Public offers for one company's records have a middle half of $174K to $693K per company, and the median of disclosed sales is $28.3K per company. In one broker's illustration, 2.1 million Gmail messages were valued at $41,000 within a larger package. The valuation gives a range for yours.

Should we let a buyer connect an app to our Gmail or Outlook? No. Google's API policy bars apps from using Workspace data to train AI models, Microsoft's bars them from building copies of API data, and buyers' counsel want each system exported by its own administrator with the vendor's native tool. Export it yourself.

Our company has closed. Can we still sell its mail? Often yes, if the mail still exists and the right person signs: in Delaware a dissolved corporation continues for three years to wind up, and in bankruptcy the trustee acts. See the estates page.

What this means for you

Your company's mail is probably its widest record and the one that reaches furthest back. It is valuable because it holds the company's work with the outside world and its decisions in writing, from every role, over every year; buyers name it and pay more for records that include it. It is fragile because a cancelled subscription deletes it, because a departing employee's mailbox can disappear with their licence, and because it is full of people and companies who never agreed to anything.

So do four things now, before any buyer conversation. Make sure a current officer holds the top administrator role and that the mail subscription stays paid. Count your mail by mailbox, year and outside domain. Mark what stays out: counsel, HR, customers whose contracts forbid the use, counterparties under NDA. And get a range for the whole record from the valuation, so that the first offer you see is measured against the market rather than against nothing. Then export it yourself, de-identify it to a standard you chose, and put it before every buyer that fits at once. The Slack guide covers the chat that usually sits beside your mail, and the rest of the What sells section shows the other records buyers pay for.