Three questions to ask any due-diligence tool before you trust its output

First: where does the underlying data come from — was it collected with the subject's knowledge, or gathered independently? This is the single biggest determinant of both legal exposure and, often, data quality (self-reported and self-connected data tends to be more current than whatever a scraper last indexed).

Second: is the scoring methodology explained, or is it a black-box number? A tool that can walk you through exactly why a founder scored where they did on each factor is one you can actually defend a decision with, internally or to the founder themselves.

Third: what happens if the finding is wrong, or the subject disputes it? A process with no mechanism for the subject to see, correct, or contest what was gathered is a process with no error correction at all — which matters, because every OSINT and scoring system makes mistakes sometimes.

Consent, explainability, and correctability aren't just nice-to-haves. They're the three things that separate a diligence tool built to last from one that's one bad news cycle away from becoming a liability.

More from the blog