# Identify valuable proprietary data before making a referral

Canonical: https://refer.sourcex.si/guides/how-to-identify-valuable-operational-data-before-making-a-referral
Last updated: 2026-11-03

Ask metadata-level questions about what the company documents, how much, over what period, in which systems and who owns it. Original, consistent, well-maintained, company-owned material is the strongest signal. You never need to see the content.

## The five questions
To identify companies in the U.S., Canada, the EU, the UK, Australia, New Zealand, India, Singapore, South Africa, Latin America, the UAE and Saudi Arabia (outside the U.S., Canada and the EU, at least 70% of the data must be in English) with valuable proprietary data, consider these questions. You're looking for insights into their internally generated, owned data.

1.  **What kinds of work do they document?** Procedures, projects, support, quality, finance.
2.  **How much?** Rough counts, such as hundreds of procedures or years of projects.
3.  **Over what period?** Consistent history is more useful than a single snapshot.
4.  **Where is it kept?** For example [Google Drive](/systems/google-drive), [Notion](/systems/notion) or [Salesforce](/systems/salesforce).
5.  **Who owns it?** The company, rather than its clients or vendors.

## Worksheet
| Signal | Strong | Weak |
| --- | --- | --- |
| Origin | Created in-house | Copied or generic templates |
| Consistency | Maintained for years | Sporadic |
| Structure | Organized, versioned | Scattered files |
| Rights | Clearly company-owned | Client or third-party content |

## Red flags
- Mostly personal data about customers or employees
- Material created under client contracts
- No one able to authorize a discussion about licensing the company's proprietary data

## Next step
If the signals are strong, [introduce the company](/guides/how-to-introduce-a-company-for-an-ai-data-partnership) for a potential data licensing partnership with AI developers worldwide.
