Support tickets
The customer's own words and your team's replies, kept with the status each ticket ended on.
A software company's code history and help desk tickets can be licensed to AI companies when the tickets link to the issues and code changes that fixed them, and when customer contracts allow that use. The repositories and the product stay with the owner. Customer names and keys in the code come out before a record leaves.
A model that writes code learns more by following one problem through every record it touched. When your ticket points to the issue and the issue points to the pull request, an AI company can see what the customer ran into and how your engineers solved it.
Support ticketHelp desk
A customer writes in: the CSV export stops halfway on large accounts.
BugIssue tracker
Support files a bug and links it to the ticket.
Pull requestCode repository
An engineer writes the fix. A reviewer asks for a test on a large file before approving it.
ReleaseCode repository
Version 4.2 ships with the fix, and its release notes name the bug.
ReplyHelp deskOutcome
Support tells the customer the export works now and sets the ticket to solved.
Support ticketHelp desk
A customer writes in: the CSV export stops halfway on large accounts.
BugIssue tracker
Support files a bug and links it to the ticket.
Pull requestCode repository
An engineer writes the fix. A reviewer asks for a test on a large file before approving it.
ReleaseCode repository
Version 4.2 ships with the fix, and its release notes name the bug.
ReplyHelp deskOutcome
Support tells the customer the export works now and sets the ticket to solved.
Your help desk and your repositories hold layers of the same work, joined wherever someone pasted a link between them.
The customer's own words and your team's replies, kept with the status each ticket ended on.
Bug reports and feature requests, with the links that tie one to the work it blocked or the report it repeated.
Every commit and pull request, with the review comments that asked why a line changed. Old branches keep the attempts that never merged.
Design notes in the team wiki, and the engineering channels where someone worked out a fix before it reached a ticket.
We raise each of these with you early. No record is prepared until you agree how it's handled.
Tickets and server logs hold your customers' names and email addresses. They come out before a ticket or log line leaves your business, and you approve the cleaned tickets first.
A password deleted from the code last year still sits in the commit history. Old commits are checked for passwords and access keys before any code is copied.
Libraries your team pulled in from open-source projects carry licenses their authors chose. That code can be left out, so the license covers what your own engineers wrote.
Customer contracts and data processing agreements set what you may do with the data customers send through your product. Your lawyer reads them first, and anything they limit stays out.
A ticket linked to its issue and pull request shows a model the whole repair. If your agents already link tickets to issues from the help desk, those links count.
Comments on a pull request show why a change was made and where reviewers pushed back before it merged.
Bugs marked as duplicates or as blocking other work show how your team chose what to fix first.
One product's repositories or a fixed run of years can make up the whole license. The rest of your code history never enters the copy.
Pick a range for your head count and for how many years your repositories go back.
Show us where your tickets and code live, and we'll explain what a license of them covers.
If you go ahead, your engineering lead describes the repositories in more detail.
We match your code history to AI companies teaching models to write software. Any company we bring you is vetted before it makes an offer.
Any offer gets explained to you clause by clause, and the signed license fixes your payment dates.
Yes, when the records connect. A ticket that links to the issue and the pull request behind it shows how your team solved a real customer problem, and AI companies license that kind of history to teach their models to write and fix code. You keep the code and the business, and the license limits how the copy gets used.
Start with the contract and the data processing agreement each customer signed. Some limit what you may do with anything a customer sends, which covers its tickets and its logs. Anything a contract limits is left out until that customer signs off in writing, and customer names come out of every ticket that goes in.
Open-source libraries carry licenses their authors chose, and those terms come with the code into your repositories. The usual scope leaves third-party code out, so the license covers the code your own engineers wrote, along with its commits and review comments.
Every tool sets its own terms for each plan, and they get revised. Read the current terms for your plan before anyone runs an export. The first call ends with a list of every system your records sit in, so you know which vendors' terms to pull.
Yes. A license covers a copy of the records you pick, so your repositories and help desk keep working as they do today. You can license one product's history or a set range of years and leave everything else where it is.
Names come out of commits and review comments as well as tickets, so a reader of the licensed copy sees a tag like Engineer 4 where a person's name was. Tell your team before anything goes, since the code history holds years of their work. The AI company sees your company's name only if you agree to it.
Answers about your team and your repositories show if the records fit.
Prefer to write? Email hello@sixonepartners.com.