SourceX Releases Guide on Licensing Company Data for AI

SourceX, a company based in Los Angeles, has released a guide detailing the types of business records it licenses for AI training and evaluation. According to the release, SourceX is focused on helping companies understand what internal data could be valuable for AI development. The guide outlines five categories of records that qualify for licensing: support conversations, business workflows, documents and knowledge, sales and CRM records, and engineering history.

Support conversations include customer problems, responses, and resolutions. Business workflows cover tasks, decisions, and outcomes. Documents and knowledge encompass the internal materials teams use to get work done. Sales and CRM records track account and deal movements over time. Engineering history includes tickets and code review threads.

These records are typically found in support, CRM, document, and collaboration tools. The exact scope and export method are confirmed during the scoping process, and the supplier decides which systems, records, and date ranges are included. The company emphasizes that the data is de-identified through approved processing routes before it is used or shared, and a diligence packet records its source, rights, preparation, and permitted use.

SourceX is targeting knowledge-work companies with 50 or more full-time employees at their peak, not counting contractors. The company operates in the United States, Canada, and the EU, and considers other markets where at least 90% of the data is in English. Companies can assess whether their records fit for licensing without uploads or commitments by visiting SourceX’s website at sourcex.si.

Founded by Noah Loul, the CEO of SourceX, the company was previously part of B2B Rocket, which was acquired by Blackpearl Group in 2025. SourceX licenses business data directly from companies for AI training and evaluation, paying them for the data. Companies sign with SourceX, not with AI labs, and SourceX manages the agreement, data preparation, and payout.

The release notes that the global point-of-sale (POS) market is expected to reach USD 124.73 billion by 2033, with fixed cashier machines still holding roughly 65% of deployments. Small and mid-sized enterprises account for about 60% of POS adoption. Retailers are upgrading their systems to include more advanced technology, with SourceX positioning its POS terminal family for retailers that need stable supply, manufacturing depth, and ODM-level customization.

SourceX offers a one-time fee per agreed dataset, with an upper bound of USD 4 million for the largest enterprise datasets. The company aims to provide a simple, controlled way for businesses to license their internal data, which is seen as a valuable resource for AI developers.