Trust Score

A Trust Score is a single, evidence-based number that summarizes how much verified, machine-readable trust a business has accumulated on Trusgent (准策). It is the headline signal on every Trusgent ID and Agent-ready Trust Passport, designed so that both human buyers and autonomous AI agents can size up a counterparty in seconds without scraping the open web or relying on self-promotion.

The Trust Score answers a narrow but important question: based on the structured evidence this business has actually submitted and that Trusgent has actually verified, how confident can a counterparty be that it is who it says it is and has done what it claims to have done? It is deliberately not a popularity ranking, a paid placement, or an opinion poll. Every point traces back to a verifiable record.

What the Trust Score summarizes

The score is a weighted summary of evidence-based factors. None of them are about marketing spend or self-description; each reflects something Trusgent can check.

•Identity completeness — How fully the business has populated its Trusgent ID and Trust Passport: legal entity details, jurisdiction, contact channels, business scope, and structured profile fields. A sparse profile carries less weight than a complete one because there is less to evaluate.

•Verification level — The depth of identity and credential verification that has been confirmed: domain and email control, registered-entity checks, and higher-assurance verifications where available. A claim that has been independently verified counts for far more than the same claim left unverified.

•Approved proof files — Documents and artifacts the business uploaded as evidence (certifications, licenses, references, registrations) that passed Trusgent review. Pending or rejected files do not lift the score; only approved proof does.

•Verified transaction reviews — Reviews tied to real, confirmed interactions rather than anonymous comments. Because each review is anchored to a verified transaction, it cannot be manufactured at scale, and it reflects actual counterparty experience.

•Unresolved disputes — Open disputes that have not been settled or addressed weigh against the score. Crucially, the existence of a dispute is not automatically fatal; what matters is whether it remains unresolved and how the business engages with it.

•Information freshness and staleness — Trust decays. Verifications, proof files, and profile information that have not been refreshed within expected windows lose weight over time, so a score reflects current standing rather than a one-time effort years ago.

•Manual adjustments recorded in audit logs — In specific cases (for example, confirmed abuse, fraud, or a successful appeal), Trusgent may apply a manual adjustment. Every such adjustment is logged in an auditable trail, so it is never an invisible thumb on the scale.

What the Trust Score is not

•It cannot be bought. No paid plan, subscription tier, ad slot, sponsorship, or featured placement raises a Trust Score. Paid plans may unlock capacity, tooling, or visibility features, but they do not move the number. This separation is foundational: if money could buy trust, the score would be worthless as a signal.

•It is not a black-box ranking. The factors above are disclosed, and the evidence behind a score is visible on the Passport. Trusgent favors transparency over opaque ranking.

•It is not a guarantee. A high score means strong verified evidence to date; it is not a promise of future behavior or a substitute for due diligence.

How the Trust Score updates over time

The score is dynamic, not a one-time grade. It recalculates as the underlying evidence changes:

•Completing verification, adding profile detail, or having a proof file approved tends to raise it.

•Earning verified transaction reviews builds it gradually and durably.

•Letting verifications or proof go stale causes gradual decay, nudging businesses to keep records current.

•Opening a dispute, or leaving one unresolved, lowers it until the matter is addressed; resolving the dispute restores standing.

•Manual, audit-logged adjustments apply in defined situations and are reversible on appeal.

Because freshness matters, a business that verified once and went quiet will not retain a top score indefinitely. The system rewards sustained, current, evidence-backed behavior.

How humans and AI agents use it

The Trust Score is built to be read by two very different kinds of users.

•Humans use it as a fast first filter — a way to shortlist counterparties before deeper diligence, read alongside the underlying proof and reviews rather than in place of them.

•AI agents consume it through the Trust API as a structured signal during agent-to-agent matching and negotiation. An agent can require a minimum Trust Score before proceeding, combine it with the Agent-readiness Score (which measures how well a business can transact machine-to-machine), and surface its reasoning to a human.

Critically, Trusgent's principle of human control means the score informs decisions but does not make them autonomously. An agent may filter or rank using the score; a person still approves the consequential action.

Transparency and limits

A Trust Score is guidance, not a verdict. It compresses a lot of evidence into one number, and compression always loses detail. The responsible way to use it is to treat it as the entry point to the Passport — check the verification level, read the approved proof, scan verified reviews, and note any unresolved disputes — and then let a human make the final call. The score points you to the evidence; it does not replace your judgment.

Related terms

•Trusgent ID — the business's verifiable, machine-readable identity.

•Trust Passport — the agent-ready profile the score sits on.

•Agent-readiness Score — the complementary measure of machine-to-machine transaction capability.

•Proof-based reviews — reviews anchored to verified transactions.

•Trust Badges — visual marks for specific verified attributes.

•Trust API — the interface through which agents read the score and underlying evidence.

FAQ