A score arrives before the relationship does
A procurement team is asked to clear a supplier it has never transacted with. There is no invoice history and no dispute record to reason from. The decision has to be made anyway, so the team buys a number: a risk score computed elsewhere, from records the buyer does not hold, about conduct the buyer did not observe.
Portability is the point of such a number, and it is also what detaches the number from any particular relationship. The same figure reads the same way to a buyer with ten years of history with that supplier and to a buyer with none. Reading it is an act of the relying party upon data assembled by a provider, and the party being described is not a participant in the reading.
That is the property the filed architecture is built against. It requires the party relied upon to perform an affirmative act and to pay for it in the units of its own capacity to act. The specification frames this as an inversion of the direction in which cost and exposure conventionally flow: in conventional delegation and capability-attenuation architectures the cost falls upon the relying party and the party relied upon is protected, whereas here the metered consequence falls upon the party such an architecture protects.
Dun & Bradstreet, described in its own terms
Dun & Bradstreet is a commercial business information company. As publicly described, it maintains a database of business entities and assigns each a persistent identifier, the D-U-N-S Number, used as a stable reference for a company across procurement, credit, and registration contexts. Around that identity layer it publishes business credit and risk information: reports and analytical scores intended to help a buyer, lender, or insurer form a view of a counterparty it may not know directly. Nothing here describes how that is implemented.
The category exists for a good reason. Absent a shared identity layer, every buyer would re-derive the same facts about the same suppliers independently, and a supplier would be legible only to parties that had already dealt with it. A persistent identifier many systems agree on is genuine infrastructure, and the analytics above it are aggregation done properly: many observations, normalized, made comparable across entities that otherwise are not.
The question that category answers is an entity-level one. Given records a single buyer cannot assemble alone, what can be said about this company. The filed architecture answers a different question, and the sections below state what it requires.
Deference, counter-signature, and a budget that is spent
Chapter 9 of U.S. Provisional Application No. 64/117,812 discloses the following. A deferring semantic agent (100) holds a request to dispatch an action and finds its own dispatch-authority predicate satisfied: it is on its own record authorized to determine the matter. It nonetheless elects to execute under a recorded determination held in the lineage field (104) of a second agent, the followed agent. An agent whose predicate fails, whose authorization gate (300) stands in the withheld state (310), or which has entered the non-executing cognitive mode (302) does not defer. Its holding is a withholding, not a following.
Two tests run first. The origin-disjointness test asks whether the two agents are assignable to a common origin-equivalence class (200). Where they are, the record is appended as an attempted deference record and given no effect: it decrements no budget, enrolls no class, and appends nothing adverse to the followed agent. The retrievability-form verification then applies three ordered conjuncts to the followed determination record: continuity against the counterparty identity record (114); a lineage commitment comprising a lineage-extent value and an accumulator root; and membership of the determination in the closed set of Section 1.8 with a matching action class. That verification is of form and of retrievability and not of merits. The deferring agent does not retrieve the enumerated entries, does not re-derive the determination, and produces no determination concerning the followed agent's conduct.
On satisfaction of both, a deference record (900) is appended and transmitted, carrying a counter-signature window field (912) declared in the deferring agent's own signed policy object (112) and binding the followed agent, which obtains the window from the record itself.
The decrement. Where the followed agent elects to counter-sign, it appends a deference counter-signature (902) and by that append decrements its own authorization budget (404), being the budget that gates its own non-advisory dispatch. The schedule is declared in that agent's signed policy object (112). The declared amount must be not less than that declared for issuance of a conduct evaluation artifact (116) by that agent, must be monotonically non-decreasing in the count of counter-signatures it has appended within a declared window, and must be bounded above by a declared maximum. Two forms are disclosed, a step function and a form linear in the count; in one embodiment of the linear form the first counter-signature within the window costs one unit and the fifth costs two. The marginal cost of being followed rises with the rate of adoption.
That decrement occurs without regard to merits. No adjudication of whether the deference was well founded, or of the merits of the underlying artifact, is performed by either agent or by any other party as a condition of it. Responsive to the budget satisfying a floor declared in the followed agent's policy object, its authorization gate (300) is written to the withheld state (310) for an enumerated set of action classes and an escalation record is emitted to its principal. The budget is replenished by the procedure of Section 4.5 and by no other, so being followed spends a replenishable quantity rather than terminating the agent.
What the deferring agent gets. Not standing. Responsive to receipt of the counter-signature, and not otherwise, it retrieves a corrector-class enumeration (904): the origin-equivalence classes (200) recorded both as having corrected the followed agent in that action class and as having had those corrections thereafter conformed to by it. Both parts of the conformance filter (906) must hold over a policy-declared observation interval, and a class satisfying correction but not subsequent conformity is excluded. Each is carried by a class-defining reference, not by the followed agent's local identifier. Those classes are written into the replenishment register (402) as admissible-but-unexercised entries, and such an entry has two effects and no third: an artifact from an assignable party is admissible to the admission evaluator (120) notwithstanding no prior encounter history, and the entry is not counted in renewal while unexercised. Replenishment follows only where an admissible artifact later arrives and exercises the entry, recorded in an exercised transition record (908), and then only where the class is absent from that register for the declared window.
Two different objects being computed
The divergence is in the object produced, not in who produces it better. A risk score is a statement about an entity, computed from records held by the provider and portable to any reader. The disclosed mechanism produces no such statement. Its specification recites that no rating, weight, score, endorsement, vouch, or standing quantity of the followed agent is retrieved, admitted, or consumed, that nothing transmitted describes that agent's standing, and that the deferring agent acquires no credibility with the enumerated parties and receives no rating from them. What is enrolled is a set of addresses that may now be heard from.
The conversion bar (502) makes that separation structural. A deference abstention entry is not converted at any stage into a scalar value, a default value, or an operand of a threshold comparison; no failure count or suspicion quantity takes it as an input; and that the deferring agent did not determine the matter for itself is not treated as a default judgment against it. A declined counter-signature is recorded as a not-determinable determination (126) naming the counter-signature as an unavailable input, and is not treated as a rejected determination (124).
The two are complementary. A portable score compresses many observations into something a stranger can read. An enrolled exposure compresses nothing; it changes which strangers an agent will listen to.
Coexistence, and what the filed architecture leaves open
In a real procurement stack the two sit at different layers. Entity-level identity and aggregated reference data answer the onboarding question: is this a real company, and what is generally known about it. The disclosed mechanism operates afterward, where an agent must decide whether to admit a particular counterparty's assertion into its own evaluator. Assignability of an arriving asserting party (118) to an enrolled class is evaluated upon the deferring agent's own records, without reference to a centralized registry.
Other things are left open by construction. Correctness of an adopted determination is not determined, since it is expressly not adjudicated as a condition of the decrement. The specification recites that no standing quantity, rating, weight, or count of adoptions consumable by any admission or dispatch decision is maintained, and that no count of counter-signatures is compared against a threshold. Quantities are declared in a signed policy object (112) without fixed values, making the counter-signature window, the observation interval, the budget floor, and the decrement parameters deployment choices rather than disclosed constants. Where two followed agents present determinations of differing outcome class upon one conduct descriptor, the deferring agent appends a deference conflict record, adopts neither, and either determines the matter itself or withholds the dispatch. A buyer wanting a single reconciled figure will not obtain one from this layer.
Declination and non-response are handled rather than punished. Where either occurs, the deferring agent may still execute under the adopted determination, appending an ungoverned reliance record and writing a reliance-bound narrowing record that narrows its own counterparty scope object held for the followed agent by the categories its policy object (112) declares. That narrowing is expressly not a sanction, and is released upon a counter-signature within the reinstatement window.
Disclosure Scope
This article describes subject matter disclosed in U.S. Provisional Application No. 64/117,812, a pending application. Nothing here is a claim construction, an opinion of counsel, or a representation about the scope of any issued claim. Element names and reference numerals used above are those of the filed specification, and where that specification declares a quantity without fixing a value, no value has been supplied here.
References to supplier-risk scoring services are to public materials and are used for comparison only; no relationship, endorsement, or infringement is asserted. Statements about any such service describe its publicly documented purpose and category qualitatively. Nothing above is an assertion about the internal implementation of any third-party service, nor an assertion that any capability is absent from it. The comparison is between what the filed architecture requires and what a reputation, credit, or verification service generally addresses.
