Confirm that the people are correctly identified and that roles, organisations and dates are accurate. Shared surnames, stale profiles and duplicates can create convincing false paths.
Employment overlap establishes co-employment, not necessarily collaboration. Communication records show interaction, not automatically trust. Public affiliations may reveal a research angle without proving personal access.
Before an introduction, ask the proposed connector whether they know the target and consider the request appropriate. Record corrections, declines and the date of validation.
Use statuses such as documented overlap, possible connection, confirmed familiarity and approved request. Avoid converting all of them into a single "strong relationship" label.
Orbb's people and path research should be evaluated on this evidence discipline. The commercial value is a credible next step that the team can act on, not an apparently complete graph whose relationships have never been checked.