Guides · Published
Sell Slack Data to AI Buyers: What Companies Can Expect
Yes, companies can sell old Slack messages and other internal records to AI buyers. Here is how it works, what it pays, and what to check before you agree.
Short answer
Yes. Companies can license old internal records, including Slack messages, to AI companies that train AI agents on real workplace conversations. The company keeps ownership of its data and chooses what is included. Pay depends on the buyer, the size and quality of the data, and a sample review, so no fixed price applies.
What buyers want from Slack data
AI labs and data companies need examples of how people really work: how support issues get resolved, how projects move, how decisions get made. Slack threads, support tickets, CRM notes, docs and code history can show that. Ordinary workplace records are a different kind of training material from public web text.
The market is new and moving fast. Reporting by Gizmodo describes shuttered startups selling old Slack chats and email archives to AI companies. According to the same report, Forbes found that SimpleClosure had processed nearly 100 of these deals over the past year, with payouts from $10,000 to $100,000 per company. Those were mostly companies that were closing down, so a live business may see different terms.
How much a company can get
No one can give you a reliable number before a buyer sees a sample. Buyers' own pages advertise company-data deals from about $10,000 to over $1 million, and most of those offers go to companies with 30 or more employees. The figure you actually get depends on:
- how much usable history you have
- how clean and complete the records are
- which tools they come from
- how much a specific buyer needs that kind of data
- what you choose to include
Be wary of any offer that quotes a number before it has looked at your data.
How a deal works
A typical path looks like this:
- Matching. You describe your company: size, industry, tools and how many years of records you have.
- Sample review. A buyer reviews a sample, usually after names and identifiers are removed.
- Offer. You see a real range and decide whether to continue.
- Contract and payment. Terms are signed, then payment follows. Plan on one to three months from the first call to payment.
At each step you can say no. You also decide what goes in the package and which buyers you meet.
Comparison: what to check before you agree
| Question | What a good answer looks like |
|---|---|
| Who owns the data? | You keep ownership. The buyer gets a license to a copy. |
| Who is named in the records? | Names, contact details and IDs are removed before sharing. |
| What is included? | You choose the channels, tools and date ranges. |
| Who pays the matching fee? | The buyer pays. You should not pay to be matched. |
| When do you get paid? | A specific timeline in the contract, not a vague promise. |
| Who reviews the legal terms? | Your own lawyer, before signing. |
Slack data specifically
Slack is a common source because it holds years of messages between staff, clients and vendors. Slack holds about 18% of the business communication market, while Microsoft Teams holds 41%, according to Embryo. So many companies with long Slack histories will also have records in other tools.
Slack messages are not all the same. Customer names, personal details, health information and private HR matters need extra care. Some workspaces also contain material that belongs to a client, not to you. Check that before you include anything.
Email is often part of the same conversation. Embryo also reports that the average professional receives 120 emails per week, so an email archive can add a lot of volume. Treat email the same way as Slack: review it before it leaves your control.
Is it allowed?
Whether a specific sale is allowed depends on your contracts, your employees' policies, your industry and where you operate. Records that belong to clients, or that hold personal or regulated information, need a legal review before anything is shared. Ask your lawyer. Smelted does not give legal or compliance approval.
As a rule, check three things first: who owns the records, what privacy notices and employee policies say, and whether any client agreement limits reuse.
Who this fits
The best fit is a company with 50 to 500 employees, founded in 2022 or earlier, with several years of records in tools like Zendesk, Jira or GitHub, Salesforce or HubSpot, Bullhorn, AutoCAD or Revit, or Benchling. Core industries include staffing and recruiting, engineering and architecture, B2B software, and science labs or biotech R&D.
Smaller companies can still ask. Buyers who take companies under 30 employees generally pay less, so an honest answer is that the offers will be smaller. Those cases are better handled on a call.
Some industries are not a fit. These include law firms, healthcare and medical billing, accounting and tax, call centers, IT managed-service providers, education, government and defense. Physical books and published works are also out of scope, because the copyright is not the company's to sell.
How to start with Smelted
Smelted is a free matching service. It finds which buyers fit your company, gives a ballpark range, and introduces you to two or three matching buyers. The buyer pays Smelted's fee. Your company pays nothing.
You can book a 15-minute call with Chris Irving at Smelted, or call 954-682-7560. For a ballpark range, use the one- to two-minute form at Smelted.
Have these ready when you call: your employee count, the tools you use and how many years of records they hold, whether the data is yours or your clients', and whether you have already talked to any other data buyer.
FAQ
How much money do companies get for selling their data?
There is no fixed price. Buyers advertise offers from about $10,000 to over $1 million, mostly for companies with 30 or more employees. A real figure comes only after a buyer reviews a sample of your records.
How can I sell my company data?
Start by listing which tools hold your records and how far back they go. Then check who owns the data and what your client contracts allow. Next, speak with a matching service or a buyer. Expect a sample review before any offer.
Can you sell data to AI companies?
Some companies do. AI labs and data companies license internal business records to train AI agents. The sale should follow your contracts, privacy rules and employee policies, and it should be reviewed by a lawyer first.
Is it illegal for companies to sell your data?
It depends on what the data is, who it belongs to, and what promises were made to the people in it. Personal data, client-owned material and regulated information carry the most risk. Ask a lawyer about your specific records before any sale.