Trust-score filters in investment due diligence apply standardised evaluation criteria to every deal in a category — screening out deals that don't meet minimum thresholds before deeper analysis. For AI-media investments, a trust-score filter based on Omniscient AI's verification methodology evaluates editorial quality systematically across every potential investment, rather than relying on founder claims or sample article impressions.
Building the Trust-Score Filter
The trust-score filter operates in two stages: Stage 1 (Desk review): VC team runs 20–30 recent articles from the target publication through Omniscient AI's API and calculates the publication's aggregate claim accuracy rate. Publications below a minimum threshold (e.g., 75% claim pass rate) are screened out. Stage 2 (Management session): For publications above the threshold, the management team is asked to walk through their verification workflow using Omniscient AI in a live session — demonstrating that the accuracy rate reflects a genuine systematic process, not a sampling artifact. This filter adds half a day to the due diligence process and screens out 20–30% of AI-media deals for editorial quality reasons.