{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp12_substrate_denial72_20260805","cell_id":"authority_legitimacy_and_consent_foundations__accounting_auditing","arm":"ORDINARY_MAX","candidate_id":"authority_legitimacy_and_consent_foundations__accounting_auditing__ORDINARY_MAX","proposal_index":1,"version":0,"title":"A Bounded, Appealable Mandate for Anomaly-Triggered Accounts-Payable Holds","problem":"In an organization that screens invoices for anomalies, a software flag or analyst concern can become a de facto payment hold even when no one can identify who is authorized to impose, extend, or release the hold; what evidence and duration are permitted; or how the accounts-payable team, budget owner, employee, or supplier can challenge it. System permissions, fraud expertise, and managerial seniority are treated as interchangeable sources of authority, although each supports a different and limited role.","actors":["Chief accounting officer or corporate controller","Accounts-payable analysts and managers","Fraud or compliance analysts","Internal audit as an independent observer","Business-unit budget owners","Treasury personnel","Employees seeking reimbursement","Suppliers awaiting payment","Independent hold-review officer","Financial-system and anomaly-model owners"],"observable_state":"The candidate applies when closed-case records show anomaly flags followed by payment-status changes without a named human decision-maker, cited mandate, decision-specific reason, expiry time, or review route; holds are extended through email or queue reassignment; similarly situated actors disagree about who may release them; or disputed holds are resolved through seniority and private escalation rather than a recorded rule. These observations must be distinguished from ordinary processing delay and from a technically erroneous flag that existing authority handles unambiguously.","consequence":"A valid invoice may remain unpaid without a defensible control decision, while an unsupported hold may be lifted through an undocumented override. Either path can impair payment-term compliance, cutoff and liability records, supplier or employee treatment, fraud-control integrity, and the auditability of why cash was or was not released.","affected_objective":"Process valid obligations according to authorized terms while controlling fraud and accounting error through decisions that are timely, traceable, consistently bounded, and open to protected review.","intervention":"Create an Accounts-Payable Exception-Hold Mandate approved by the existing controller authority. The mandate makes anomaly systems advisory only; delegates temporary hold power to specifically qualified human officers; limits that power by transaction class, evidence category, amount, duration, renewal conditions, and prohibited uses; and states that a hold does not establish fraud, employee fault, or supplier misconduct. Each hold receives a protected reason record naming the officer, authority clause, evidence category, expiry, next fact needed, and review route. An affected internal owner or payee-facing liaison may provide evidence or request review without receiving protected detection details. A reviewer outside the originating decision chain can affirm, narrow, replace, or release the hold. Supplier or employee silence is never labeled consent; contractual authorization counts only when it is explicit and scoped, with notice and review providing the legitimacy connection otherwise.","structural_mapping":[{"archetype_element":"Binding decision requiring legitimate authority","domain_realization":"Temporarily preventing release of an otherwise payable invoice binds accounts-payable staff, treasury, the budget owner, and the payee and cannot derive authority from an anomaly score alone."},{"archetype_element":"Authority Mandate","domain_realization":"A controller-approved charter identifies the organizational payment-control authority from which temporary hold power is delegated and identifies the human roles that may exercise it."},{"archetype_element":"Authority Boundary","domain_realization":"The charter limits eligible transaction classes, evidence grounds, monetary jurisdiction, maximum duration, renewal count, escalation triggers, and consequences; it excludes ledger reclassification, disciplinary findings, contract amendment, and indefinite withholding."},{"archetype_element":"Legitimacy Basis Map","domain_realization":"The basis map separates formal payment-control delegation, domain competence, explicit contractual authorization where present, affected-party notice, representation, and independent review so that expertise or system access cannot silently substitute for mandate."},{"archetype_element":"Affected Party Recognition","domain_realization":"The process names the payee, employee, budget owner, accounts-payable operator, treasury function, and investigation function as differently affected parties, including parties who cannot realistically opt out."},{"archetype_element":"Consent Scope","domain_realization":"Only an explicit, relevant contract term may count as scoped authorization for a verification hold; bundled terms, silence, economic dependence, or merely submitting an invoice are not described as voluntary consent."},{"archetype_element":"Competence Evidence","domain_realization":"Hold officers and reviewers must have role-specific authorization and demonstrated familiarity with accounts-payable controls, evidence handling, cutoff implications, and protected investigations."},{"archetype_element":"Public Reason Record","domain_realization":"A decision record exposes the governing criterion, evidence category, duration, next action, and review route while redacting details whose disclosure would enable evasion or violate confidentiality."},{"archetype_element":"Voice Channel","domain_realization":"A business owner or payee-facing liaison can submit missing facts, identify identity or fulfillment errors, and request independent review through the case record rather than personal escalation."},{"archetype_element":"Accountability and Review Path","domain_realization":"An independent reviewer has real power to change the hold, and the controller can examine expired, jurisdictionally invalid, repeatedly renewed, or informally overridden cases."},{"archetype_element":"Conflict-of-Interest Guardrail","domain_realization":"An officer must recuse when responsible for the disputed purchase, vendor relationship, original accounting entry, or investigation outcome."},{"archetype_element":"Legitimacy Health Indicator","domain_realization":"The controller reviews the presence of ownerless holds, expired holds, jurisdiction objections, off-system overrides, incomplete reasons, review outcomes, and repeated renewals as signals of mandate drift rather than as proof of misconduct."}],"mechanism_mapping":[{"mechanism_slug":"charter_or_mandate_document","role":"Codifies the source, purpose, limits, duration, amendment conditions, and revocation conditions of temporary payment-hold authority.","counterfactual_removal":"Without the charter, qualified staff may still detect anomalies, but permissions, expertise, and seniority remain contestable substitutes for a bounded right to stop payment."},{"mechanism_slug":"decision_rights_matrix","role":"Separates who may flag, investigate, impose, renew, review, release, advise, and audit a hold; the anomaly model may only flag.","counterfactual_removal":"Without the matrix, the same system status can be interpreted as a recommendation, command, or veto by different actors, recreating ambiguous jurisdiction."},{"mechanism_slug":"public_reason_giving_protocol","role":"Requires a minimally sufficient, access-controlled explanation linking each hold to a charter clause, evidence category, expiry, and next action.","counterfactual_removal":"Without a reason protocol, affected actors cannot distinguish a properly authorized hold from queue inertia, and reviewers cannot test whether the officer stayed within mandate."},{"mechanism_slug":"appeal_or_review_forum","role":"Provides a protected, outcome-changing review by someone outside the originating decision chain, with authority to affirm, narrow, replace, or release the hold.","counterfactual_removal":"Without an outcome-changing review, objections return to private escalation, and the authority remains non-corrigible even if its scope is documented."},{"mechanism_slug":"participatory_consultation_process","role":"Before adoption, obtains structured input from accounts payable, treasury, budget owners, investigation staff, and payee-facing personnel about consequences, notice, evidence access, and workable boundaries.","counterfactual_removal":"Without consultation, the mandate may overlook parties who bear operational costs or create a review channel that exists on paper but cannot be used safely."}],"causal_chain":["An anomaly system or analyst identifies a transaction requiring examination but does not itself change payment status.","A designated hold officer checks whether the transaction, evidence category, amount, and requested duration fall within the controller-approved mandate.","If the criteria are met, the officer records the authority clause, protected reason, expiry, next fact needed, notice recipient, and review route before imposing the temporary hold; otherwise the invoice remains in the ordinary process or is escalated to an authority that actually has jurisdiction.","The affected internal owner or payee-facing liaison can provide corrective evidence or request review without being forced to accept an accusation or receive protected detection logic.","An independent reviewer can affirm, narrow, replace, or release the hold and must record the basis for doing so.","Expiry and renewal rules prevent a temporary control from silently becoming indefinite authority, while exception logs expose boundary breaches, conflicts, and off-system overrides to the controller.","If the causal hypothesis is correct, contested cases are converted from struggles over hidden standing into reviewable questions about mandate, evidence, scope, and remedy, producing a traceable basis for either withholding or releasing payment."],"baseline":"The baseline is an anomaly queue in which a model owner, fraud analyst, accounts-payable manager, or senior executive can cause or prolong a hold through system permissions, custom, email, or escalation. The system records a status change but may not record the human authority, governing criterion, expiry, affected-party route, or reason for an override. Model tuning and analyst judgment occur, but the distinction between recommending investigation and authorizing nonpayment remains implicit.","nearest_rivals":["Improve anomaly-model calibration and vendor-master data. This can address erroneous flags and identity mismatches, and should be preferred when detection quality fully explains the cases; it does not establish who may bind others after a technically defensible flag.","Add dual approval or segregation of duties for payment-status changes. This constrains unilateral action but can leave two approvers exercising the same undefined, unbounded authority with no affected-party review.","Adopt case-management service levels and automatic escalation. This addresses queue age and ownership but can accelerate an illegitimate decision or escalate it to seniority without clarifying jurisdiction.","Centralize all exception decisions in a specialist fraud or accounts-payable team. Centralization can improve consistency and competence, but expertise alone does not authorize unlimited duration, collateral consequences, or denial of review.","Use post-payment continuous auditing instead of prepayment holds. This avoids some withholding decisions but shifts exposure to recovery after payment; it is a serious substitute where transactions are reversible and recovery risk is acceptable.","Require controller or audit-committee approval for every hold. This supplies formal hierarchy but may overload senior forums and still lacks case-level boundaries, reasons, representation, and independent correction."],"remaining_contrastive_claim":"The remaining claim is conditional and structural: when technically defensible anomaly flags still produce conflict because advisory expertise or system access is mistaken for unbounded payment authority, a human-held, controller-mandated, time-limited decision right joined to protected reasons and outcome-changing review addresses that legitimacy defect in a way that detector tuning, service levels, centralization, or dual approval alone do not. There is no residual claim if existing authority is already legible and corrigible or if data and detection defects fully explain the observed cases.","authority_safety":{"decision_authority":"Only the chief accounting officer or controller who already owns payment-control policy may approve the mandate or a live pilot. After approval, only named and qualified hold officers may impose or renew holds within its limits. A reviewer outside the originating decision chain may modify or release them. The anomaly model, research team, internal audit observer, and affected business manager receive no new payment authority.","authorized_first_step":"The controller and data owner may authorize a read-only retrospective review and tabletop simulation using already closed cases. Participants may map existing authority, reconstruct decisions, and test a draft mandate, but the draft has no operational force.","excluded_actions":["No live invoice may be held, released, reprioritized, or reclassified under the draft.","No accounting record, vendor master record, reimbursement status, or investigation disposition may be changed.","No automated system may receive payment-decision authority.","No employee or supplier may be contacted or characterized as suspicious for the evidence exercise.","No retroactive discipline, performance action, recovery demand, or adverse vendor action may be based on the review.","No data may be accessed beyond participants' existing authorization, and protected fraud logic may not be placed in broadly visible records.","No silence, boilerplate acceptance, or economic dependence may be recorded as consent."],"halt_rollback":"Halt if the exercise requires unauthorized data, exposes protected investigation methods, reveals an unmanaged participant conflict, or begins influencing a live payment. Mark the draft non-operative, return all cases to existing authority paths, restrict working materials under existing retention rules, and document why the evidence exercise stopped. Because the first step is retrospective and read-only, rollback consists of withdrawing the draft and ending access rather than reversing transactions."},"negative_tests":{"strongest_counterevidence":"The strongest counterevidence would be records showing that current policy already identifies a qualified human decision-maker, bounded grounds and duration, meaningful notice, conflict rules, and an independent review that can change outcomes, while every disputed or prolonged case is attributable to missing data, model error, or ordinary processing capacity rather than contested authority.","problem_falsifier":"The proposed problem is falsified for the sampled setting if each reviewed hold can be traced to an operative mandate and authorized officer, all parties agree on jurisdiction and review rights, and no hold duration, renewal, override, or dispute depended on ambiguity about who could decide or why.","intervention_falsifier":"The intervention mechanism is falsified if, in blinded tabletop cases, authorized participants still cannot determine who may decide, reviewers cannot alter an originating decision, or the required reason record cannot support challenge without disclosing information that must remain protected. It is also falsified if the draft merely restates an already operative process and changes none of the authority relationships implicated by the cases.","risks":["A formal charter could legitimize poor anomaly detection instead of correcting it.","Extra review steps could delay valid payments or urgent fraud containment.","Reason records could expose confidential investigations, personal data, or detection logic.","Staff could treat contract boilerplate as consent despite weak comprehension or bargaining power.","Specialist reviewers or consultation channels could be captured by dominant business units or investigation staff.","Appeals could become symbolic if reviewers lack independence, information, or power to change outcomes.","Rigid boundaries could omit a genuinely urgent case or encourage actors to relabel decisions to evade the mandate.","Metrics could be gamed by resolving cases off-system or discouraging review requests.","The separation between a temporary hold and an accusation could collapse in practice, causing reputational or employment harm.","The charter could diffuse responsibility if flagging, holding, reviewing, and releasing roles assume another role owns the final outcome."]},"next_evidence_step":"With controller and data-owner approval, select no more than 30 closed anomaly-flagged invoices from one legal entity and one completed accounting period, including available cases with extensions, overrides, or recorded disputes and a comparison set without them. For each case, reconstruct who flagged, imposed, extended, reviewed, and released the hold; the asserted authority source; evidence category; duration; notice; conflict status; and whether any review could change the outcome. Before inspection, define authority-friction indicators as an unknown decision owner, disputed jurisdiction, uncited mandate, extension beyond a stated limit, private escalation, or unavailable outcome-changing review. Then run a tabletop application of the draft charter on at most six de-identified closed cases with an accounts-payable officer, controller delegate, payee-facing representative, and independent reviewer. Record disagreements and stop under the stated safeguards. This step tests whether the problem is present and whether the proposed mechanisms are interpretable; it does not alter live payments or estimate an effect size.","prior_art_status":"UNSEARCHED","diversity_from_prior_proposals":"Not assessed because runtime isolation prohibited inspection of other proposals. This candidate is specifically instantiated as legitimacy infrastructure for human authority over anomaly-triggered accounts-payable holds.","revision_record":{"parent_version":null,"progress_targets_addressed":["Initial version; no prior proposal feedback or revision target was supplied."],"conceptual_changes":["Formulated the domain problem as a separation between anomaly detection competence and legitimate authority to bind payment actors.","Used mandate, boundary, affected-party connection, competence, reason-giving, conflict safeguards, and corrigible review as jointly necessary structural elements.","Explicitly rejected fictional consent and treated notice, representation, and review as the affected-party connection where voluntary opt-out is unavailable."],"operational_changes":["Specified a controller-approved hold charter, decision-rights matrix, protected reason record, independent review, expiry rules, and recusal conditions.","Limited the first step to a read-only retrospective sample and de-identified tabletop exercise."],"evidence_changes":["No external evidence, prior experiments, other candidates, or prior-art search was used.","Added observable indicators and separate problem and intervention falsifiers for a bounded internal evidence test."],"claim_changes":["Made no claim of novelty, prevalence, demand, or effect size.","Restricted the contrastive claim to settings where authority ambiguity remains after detection and data-quality explanations are tested."]}}