Licensing Operational Data from SaaS Companies: Jira, GitHub & Support Tickets

A SaaS company's most licensable data is its internal record of building and running software: Jira tickets, Confluence pages, internal Slack, incident postmortems and planning docs. Customer data you process under contract, including most support tickets, is different. Your data processing agreements likely limit it to providing the service, so scope it out or clear it with counsel.

Why do AI labs want software-company workflow data?

AI developers are training agents to operate the same tools your team uses every day. Scale AI says nearly half of its new data training projects involve reinforcement learning environments, and Turing sells simulated clones of Jira, Salesforce and Zendesk for agent training. Real records of how people actually used those tools, ticket by ticket, are the input those environments imitate.

Micro1's chief revenue officer told Business Insider that buyers increasingly want "long-horizon" data that shows how work gets done end to end, and that if a company won't include Slack, CRM or email, the rest of the corpus is worth less. A software company's chain from customer bug report to Jira ticket to code fix to release note is exactly that kind of sequence. Dan Novaes used a version of it as his example on The LFG Show: a bug ships, finance has double-billed customers to refund, and the fix gets committed.

Does your company qualify?

Software companies fall in Mode's general category, not the lower legal or accounting tiers.

Offshore engineering teams matter here. Mode says mostly U.S.-based teams are the strongest fit, so a 40-person company with 25 engineers abroad should expect questions. Confirm the current threshold during intake.

What is in scope and what isn't?

Mode's program page lists Jira, Confluence, Slack, Notion, Zendesk, Salesforce, HubSpot, Google Drive and Microsoft among its sources, plus "+1K more." It does not name GitHub or GitLab. Sort your systems by who owns the content inside them.

The distinction that matters is controller versus processor. Your own Jira is your record. The data sitting in your product, and much of what flows into your support desk, belongs to your customers and their users. You hold it to provide a service.

Why are support tickets and customer data the hard part?

Most B2B SaaS contracts include a data processing agreement or equivalent terms that restrict use of customer data to providing and supporting the service. Under the CCPA, a service provider is generally barred from using personal information it receives for other purposes; under the GDPR, a processor acts only on the controller's documented instructions. Licensing support tickets to a data buyer is a new purpose your customers did not authorize. That remains true even if names are later transformed. The question is whether you may disclose the data at all, not just what the buyer finally sees. Read your DPAs, master subscription agreements and security commitments, and have counsel decide. If you can't pull a clean subset, leave support data out.

Your SOC 2 report and security questionnaires may also describe data-handling commitments. A customer reviewing your next audit may ask where its data went.

Employee data is the other layer. California employees' personal information has been fully covered by the CCPA since January 1, 2023, according to the California Privacy Protection Agency. Emory law professor Ifeoma Ajunwa told Business Insider that whether employees' consent extends to the sale of their data is "an ethically gray area." Engineers who wrote candid postmortems did not write them for an outside buyer. Tell them before you sign.

What about licensing source code?

Treat code as its own transaction, not a line item in a workflow deal. Three issues come first:

• Rights. Repositories mix your code with open-source and vendored third-party code under licenses you have to respect, plus contractor code whose assignment you should confirm. Business Insider reported that a Surge AI engineer was offered a large company's stolen codebase, which is why buyers increasingly check provenance.
• Secrets and customer specifics. Repos and commit history hold API keys, credentials and customer-specific configuration. Rotate and scrub before anything leaves.
• Exclusivity and competitive value. Business Insider reported lower demand for a codebase a rival lab already has. Exclusivity raises price but limits future deals. Your code is also your product; consider what an AI-assisted competitor could do with it.

Ask Mode directly whether repositories are in scope and on what terms.

What makes software workflow data more or less valuable?

Mode says it reviews "the systems and dates in scope, years of history, record volume and quality, and ownership and access terms" before making a fixed offer. Read against what buyers have told reporters, that list suggests what moves the number for a software company.

Two practical points follow. First, a company that excludes support data and shared customer channels should expect a lower number than its headcount suggests; that is the cost of staying inside your contracts. Second, before you grant exclusivity, decide whether you might want to license the same records again, or license source code separately, later. Exclusivity is a pricing lever for you as much as for the buyer.

What could a software company be paid?

No public figure is specific to SaaS companies. The attributed ranges:

Mode says its calculator "is illustrative, not a guaranteed offer." The estimate below applies Mode's general-category minimum and published range to your headcount. It can't see which systems you exclude, so treat it as a headcount-based starting point, not a forecast. See our valuation guide and methodology.

How does the process run?

Mode's release describes an initial estimate, a 14-day nonbinding letter of intent with one 7-day extension at Mode's request, then 15–30 days of collection and de-identification after you accept the final offer, with payment "expected within approximately 60 days." Mode's program FAQ now says about three months from first conversation to payment. You approve systems and date ranges before export, and your team signs in to approved systems; no passwords are handed over. Our process guide has the detail.

What should a CTO and general counsel check?

• A system-by-system scope listing workspaces, projects and channels, with date ranges.
• Customer contracts and DPAs for any system holding customer data. Default to exclusion.
• Secrets scan of every in-scope wiki, ticket and repo.
• Filtering evidence. Mode's release says it built its own PII-removal tools after other vendors would not let it audit theirs. Ask for a description and sample output from your own data. Do not accept claims that data is fully anonymous.
• Exclusivity terms, and how they affect future deals with other buyers. Compare options in our broker comparison.
• Investor and lender consents if your financing documents restrict asset dispositions or IP licenses.
• Staff notice before export.

Who are we?

DataLicensingAudit is an independent referral partner of Mode. If your company signs after we introduce it, Mode may pay us a referral fee; that does not change your payment. See our disclosure. We don't give legal advice. Talk to your counsel before you scope anything.