Which portfolio companies are not a fit for data licensing?

The short answer: rights and records rule companies out, not industry

A portfolio company cannot license its data when it lacks clean rights to the records or lacks records worth licensing. In practice that means files that mainly belong to clients, a core dataset of consumer personal data or protected health information with no licensing basis, deleted archives, a prior AI training license on the same data, or nobody able to run an export.

Run this exclusion pass before the screening conversation, not after. It saves a CEO meeting, protects the operating team's credibility, and gives the companies that pass a better-prepared introduction.

The exclusion list for a portfolio screen

Why each disqualifier matters

Client-owned records

An outsourcer's ticket queue or an agency's creative files often document the client's business more than the outsourcer's own. Licensing them without consent breaches confidentiality terms. The outsourcer's internal records, such as staffing decisions, quality reviews and its own procedures, may still be in scope.

Consumer personal data and privacy promises

A business that collected customer data under a privacy policy cannot quietly widen it. In February 2024 FTC staff warned that adopting more permissive data practices, such as using consumers' data for AI training, and telling consumers only through a surreptitious, retroactive change to terms of service or a privacy policy may be unfair or deceptive. The post is staff guidance, not a rule, but it explains why consumer-heavy datasets rarely clear.

Protected health information

Health records offered for licensing generally must be de-identified under HIPAA's standard or covered by valid authorization. HHS describes the two methods, Expert Determination and Safe Harbor, in its HIPAA de-identification guidance. A healthcare administration business whose value lies in scheduling, staffing or billing operations can still be a candidate; one whose value is the medical record itself usually is not.

Financial customer information

Consumer lenders, finance companies and other businesses covered by the FTC's Gramm-Leach-Bliley Act guidance face notice and opt-out duties before sharing customer information with certain nonaffiliated third parties. Their internal operating records, stripped of customer information, are a separate question for counsel.

Prior licenses, deletions and missing exports

These are factual checks, not legal ones. Ask whether the data was licensed before for AI training, which systems were retired and whether exports were kept, and who holds admin rights today.

What looks like a disqualifier but often is not

The qualification overview explains the positive criteria; these are the false alarms sponsors raise most.

• The company was acquired or wound down. Operating, acquired and wound-down companies can all qualify if the data still exists and someone with authority can sponsor the license.
• The company has shrunk. Headcount is measured at peak, so a business that once had 50+ full-time employees can qualify after downsizing.
• Records are spread across many systems. That is normal; strong candidates often run 10-15+ systems.
• Some records contain personal information. De-identification and redaction requirements are agreed with the company before any work begins.
• A migration left some history behind. What survived, including archived systems, may still cover several years.

The five-minute exclusion pass

Ask these before raising data licensing with a portfolio CEO. For the first six questions a yes parks the company; for the last two a no does.

• Do the records mainly describe clients' businesses under client contracts?
• Is the core dataset consumer personal data, or medical records and claims?
• Were key archives deleted, or tools cancelled without an export?
• Is a court, trustee, receiver or assignee in control without having been involved?
• Has the same data already been licensed for AI training?
• Did the company never reach 50+ full-time employees at peak (contractors excluded)?
• Would the owner consider an exclusive AI-training license for an agreed term?
• Can someone at the company run exports from the main systems today?

The network opportunity finder helps sort a wider network before this pass, and the guide to data provenance and rights documentation covers what a passing company should prepare. Questions that remain once a company passes, such as whether customer consent is needed or board approval for a data license, have their own pages.

What to tell a CEO whose company does not fit

Be direct, and leave the door open where the problem is fixable.

How exclusions fit a sponsor's referral work

Screening out early is part of the job. The companies that remain are the ones worth an introduction, and the private equity operating partners page explains that role. Partners earn 25% of the eligible platform fees SourceX actually collects from a referred company's licensing deals, capped at $100,000 per referred company and payable only after the buyer pays and SourceX receives its fee. Introducing a company that fails the baseline produces no deal and no reward, and it slows down the ones that pass.

This is general information, not legal, tax or financial advice. Confirm with your own counsel before acting.

Next step

Check the remaining companies against who qualifies, then register as a partner and introduce the ones that pass. A company can also apply directly at sourcex.si/apply.