SHIELD Trust Score.
A single 0-to-100 number that lets a bank compliance officer or a FinTech partnership lead answer the first question they always ask: can I trust the organization on the other side? The Trust Score is computed continuously from verifiable signals, shown on every profile, and logged immutably every time it changes.
What it means
Every member of TECH has a Trust Score. The number is a weighted roll-up of four independently-measured dimensions: KYB verification status, platform activity, reputation signals, and compliance posture. Each dimension contributes a sub-score that is visible on the organization's profile alongside the overall number — no black boxes.
70–84 — trusted with minor gaps (e.g., KYB in renewal window).
50–69 — monitoring zone. Shown with caution badge.
Below 50 — not recommended. Excluded from most matches by default.
How it's computed
The formula is fixed and public — we believe trust scoring should be deterministic and auditable, never a proprietary black box. The total is a weighted linear combination:
trust_score = (kyb × 0.30) + (activity × 0.25) + (reputation × 0.25) + (compliance × 0.20)Persona API-based business verification: legal entity match, beneficial ownership (UBO), sanctions screening (OFAC, EU, UN), politically exposed person (PEP) checks, and incorporation proof. A passing KYB gives 100; a pending KYB gives 60; a failed or missing KYB gives 0.
Days since last login, profile completeness, number of active team members, match response rate, and message cadence. Ghost accounts drop here fast. A fully-onboarded org with weekly engagement scores in the 90s; a dormant account trends toward zero over 90 days.
Feedback ratings from past partners, completed deals in CORTEX Pipeline, published case studies, product validation from active integrations, and absence of security complaints. Reputation is the slowest-moving component and the hardest to game — which is why it carries a quarter of the weight.
Consent records (GDPR / CCPA / FINRA archival), two-factor authentication enabled, IP allowlist configured, data retention policy set, and no unresolved security alerts. Missing any mandatory consent drops this to 0 immediately and freezes the overall score until the gap is closed.
When it changes
Trust Scores are re-calculated nightly at 05:30 UTC by a Celery Beat task, and immediately on any of the following events:
- KYB status transition (submitted → verified, or verified → expired).
- Consent accepted or revoked.
- Security alert raised or resolved by SHIELD.
- Partnership feedback recorded in CORTEX.
- Team member added, removed, or deactivated.
Every change writes a TrustScoreHistory row with the old
score, new score, the sub-scores that changed, and the triggering
event. The full history is visible to the org's admins and exported in
the SHIELD compliance audit trail.
How it's used across the app
- PRISM matching — matches are skewed toward partners with Trust Scores ≥70. Below-threshold orgs can still appear if explicitly whitelisted, but never in the auto-generated daily feed.
- Directory sorting — default sort order for the Member Directory is Trust Score descending, so the most reliable partners surface first.
- Alert triggers — a drop of 10+ points in a single day triggers a SHIELD alert to the org admins and, if the org is already in an active deal, to their counterparty.
- FINRA/SEC audit — the score is part of every compliance export. Trust Score at the time of a message, match, or deal is preserved alongside the event for retrospective review.
API surface
It lives on every member profile.
Open your own Organization Profile or any member in the Directory — the Trust Score appears in the header with all four sub-scores visible on hover.
Open SHIELD →