What business data qualifies?

Start with records of real work

Useful starting points include both the instructions a team follows and the history of how work happened. Consider which records belong together, how far back they go and which systems hold them. Examples include:

• SOPs, playbooks, checklists and internal knowledge bases.
• CRM records, pipeline history, support tickets and quality reviews.
• Project histories, task records, specifications and reports.
• Leases, rent rolls, underwriting models and property files.
• Financial statements, reconciliations, invoices and approvals.
• Technical documentation, engineering workflows, drawings and CAD.

Describe the context, not just the file count

A useful inventory names each source system, the years covered, the approximate volume and the work the records describe. Note whether files are organized and complete, and whether a project or workflow connects information across systems. DataShelf considers volume, complexity, uniqueness, quality and domain when discussing buyer interest.

Check exclusions before sharing

Keep patient or clinical records, payment card and bank account numbers, identity documents, children's data, biometric data, individual HR and payroll records, privileged legal material, credentials and export-controlled material out of submissions. Content scraped from websites or licensed from someone else is also excluded.

The sample guidance explains what to leave out before an upload. If you are unsure about a record, leave it out and ask hello@datashelf.ai.

Begin with an inventory

Identify the systems your company uses and flag known rights or confidentiality restrictions. Start the data manifest with this description. If a buyer later needs samples, we arrange them with you on a call and send a private upload link.