SourceX has unveiled a new assessment process for companies curious about the value of their business data for AI training, with a key twist: no files need to be uploaded to get an initial answer. The Los Angeles-based company's business data licensing assessment begins with a fit check consisting of a few quick questions and no commitment, according to a press release issued November 4, 2026.
The approach addresses a common hesitation among leaders of established companies who wonder whether their support histories, CRM records, and engineering tickets hold licensing value but are reluctant to export sensitive material just to find out. "Nobody should have to hand over their data just to learn whether it's worth something," said Noah Loul, founder and CEO of SourceX. "Answer a few questions and we'll give you a straight answer on fit. If it's a yes, you stay in control of every step that follows. If it's a no, you've lost nothing."
The fit check evaluates the shape of a business rather than its contents. SourceX seeks companies that had 50 or more full-time employees at peak, excluding contractors. Knowledge-work companies are the strongest fit, and operating, acquired, or wound-down businesses can all qualify. Geography also matters: the United States comes first, followed by Canada and the European Union, with several other markets eligible when at least 90% of the data is in English. Not every company qualifies, and the check is designed to say so early.
When a fit is identified, the conversation moves to scope. The company decides which systems, records, and date ranges it is willing to include, typically drawn from support, CRM, document, and collaboration tools, with the export method confirmed during scoping. A rights review then confirms the company owns those records and can license them, including any limits in its customer or vendor contracts. Only after qualification does a company sign a Master Supplier Licensing Agreement, followed by a short Dataset Deal Confirmation for each deal. Nothing is delivered until both are signed and SourceX gives the go-ahead. Large datasets can stay in the supplier's own storage or ship on encrypted drives, and data is de-identified through approved processing routes before it is used or shared.
SourceX currently runs organization-level access, private storage with short-lived links, a full audit trail, and terms locked per deal. SOC 2 Type II, ISO/IEC 27001, and ISO/IEC 27701 are all in progress and have not been issued.
The announcement matters for technology and business leaders weighing the potential of their proprietary data in the AI economy. By eliminating the need to upload files for an initial assessment, SourceX lowers the barrier for companies to explore a new revenue stream while maintaining control over sensitive information. The staged process—from fit check to scoping to rights review to signed agreements—offers a structured path for licensing data directly to an intermediary rather than to an AI lab. Companies can start the assessment at SourceX, and more information is available at sourcex.si.
