{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp09_archetype_breadth150_20260804","cell_id":"rate_limiting__accounting_auditing","arm":"BREADTH_PROBE_ONE_SHOT","candidate_id":"rate_limiting__accounting_auditing__P1","proposal_index":1,"version":0,"title":"Metered Admission of External-Audit Evidence Requests","problem":"During fieldwork, multiple audit workstreams can release valid evidence requests to the same client accounting teams faster than those teams can retrieve, review, and authorize records. The resulting burst creates a growing request backlog, interrupts close and control work, and encourages rushed, incomplete, or duplicative responses even though each request may be individually justified.","actors":["External-audit engagement manager","Audit workstream leads","Client audit liaison","Client accounting process owners","Client controller","Internal control or compliance personnel"],"observable_state":"The evidence-request log shows sharp bursts of newly assigned requests, many simultaneous requests per process owner, rising request age, repeated deadline extensions, duplicate follow-ups, and close or control tasks competing with audit-response work.","consequence":"Finite accounting-team capacity is captured by whichever audit workstream releases requests first or follows up most aggressively, increasing response latency and the likelihood of incomplete evidence, avoidable rework, and disruption of time-critical accounting duties.","affected_objective":"Obtain sufficient, timely audit evidence while preserving reliable financial-close and control execution and allocating client response capacity transparently across audit workstreams.","intervention":"Place an admission boundary before an approved audit request becomes active client work. Use a token-bucket policy for each process-owner group: tokens refill weekly, each newly activated request consumes a token, a small burst allowance accommodates planned testing events, and unused capacity may be reassigned. Reserve a separate allowance for requests tied to imminent reporting deadlines or identified high-risk matters. Requests beyond the current allowance remain visible in a pre-admission register with an owner, reason, and proposed activation date; they are not silently queued or sent through side channels. The engagement manager and client controller review utilization and exceptions at a fixed cadence and may adjust future limits when measured capacity changes.","structural_mapping":[{"archetype_element":"Flow","domain_realization":"Audit evidence requests moving from audit workstreams into client accounting teams for action."},{"archetype_element":"Finite shared capacity","domain_realization":"Process owners' limited weekly capacity to retrieve, validate, approve, and transmit evidence while performing close and control duties."},{"archetype_element":"Measurable unit and subject","domain_realization":"One activation-weighted evidence request attributed to its audit workstream and responsible client process-owner group."},{"archetype_element":"Measurement window","domain_realization":"A rolling weekly allowance, supplemented by a monitored ceiling on simultaneous active requests."},{"archetype_element":"Admission control boundary","domain_realization":"The shared request register changes a request from proposed to active only when the relevant allowance is available or an authorized exception is recorded."},{"archetype_element":"Prioritization and fairness policy","domain_realization":"Reserved capacity protects deadline-critical and high-risk audit work, while ordinary allowances prevent one workstream from consuming all response capacity."},{"archetype_element":"Predictable excess handling","domain_realization":"Excess requests remain in a visible pre-admission register with planned activation dates instead of reaching process owners or accumulating in an undisclosed backlog."},{"archetype_element":"Observability and auditability","domain_realization":"Token use, denials, delays, exceptions, side-channel attempts, request age, and capacity changes are retained in the engagement log."}],"mechanism_mapping":[{"mechanism_slug":"token_bucket","role":"Allows a bounded number of new evidence requests per process-owner group, permits a small legitimate burst, and constrains sustained intake through a weekly refill rate.","counterfactual_removal":"Without the token bucket, approved requests can again be released in an unbounded burst, so the intervention becomes coordination guidance rather than rate limiting."},{"mechanism_slug":"priority_class_rules","role":"Protects a separately stated allowance for imminent reporting deadlines and identified high-risk matters.","counterfactual_removal":"A single undifferentiated allowance could let routine requests consume capacity needed for critical audit evidence, creating priority inversion."},{"mechanism_slug":"pre_admission_register","role":"Makes delayed demand explicit and supplies predictable handling without assigning it prematurely to client process owners.","counterfactual_removal":"Excess demand would likely move into email, informal promises, or an invisible queue, preventing governance of backlog growth and limit evasion."}],"causal_chain":["Concurrent audit workstreams generate individually valid evidence requests.","Without an intake-rate boundary, release timing rather than an explicit allocation rule determines which requests consume client accounting capacity.","Bursty activation exceeds the retrieval, review, and authorization rate of shared process owners.","Backlog, interruptions, deadline extensions, and rushed responses increase.","A token-bucket rule meters request activation by process-owner group over a weekly window while permitting a defined burst.","Priority reservations preserve capacity for deadline-critical and high-risk requests; other excess demand remains visibly unadmitted.","The active request load stays within the declared response envelope, making tradeoffs and exceptions observable instead of allowing uncontrolled intake."],"baseline":"A common project-control baseline for comparison is a shared request list with due dates and escalation meetings, but no enforceable limit on how many requests workstreams may activate for a process-owner group during a time window.","nearest_rivals":["Centralized request portal: consolidates visibility and communication but need not bound the rate at which requests become active.","Work-in-progress cap: limits simultaneous active requests but may not control rapid replacement or total admissions over a measurement window.","Audit plan scheduling: sequences expected procedures in advance but may not govern unplanned request bursts during fieldwork.","Priority triage: ranks requests after demand appears but does not itself impose a sustained admission envelope.","Additional client or audit staffing: expands or redistributes capacity rather than rationing admission to the existing capacity."],"remaining_contrastive_claim":"The candidate's distinguishing causal move is enforceable, time-windowed admission of audit evidence requests before they become client work; visibility, scheduling, or priority labels alone do not preserve the proposed bounded-intake invariant.","authority_safety":{"decision_authority":"The external-audit engagement manager owns request-release rules, while the client controller validates process-owner capacity and operational constraints. Changes affecting audit scope, required evidence, risk assessment, or statutory deadlines remain with personnel who already hold those responsibilities.","authorized_first_step":"Conduct a read-only shadow test using one completed fieldwork period: replay timestamped requests through candidate weekly token and priority rules without changing past conclusions, current assignments, audit scope, or communications.","excluded_actions":["Suppressing or deleting a required audit request","Changing audit scope, risk assessment, sample selection, or evidence-sufficiency judgments through the rate rule","Allowing client personnel to veto a request solely because its allowance is exhausted","Delaying a request past a legal, regulatory, reporting, or evidence-preservation deadline","Sending unlogged requests through email, chat, or verbal side channels","Automatically treating an unadmitted request as resolved or unnecessary","Using individual employee productivity judgments to set or enforce quotas"],"halt_rollback":"Halt the trial if a critical request would be delayed, the reconstructed register cannot distinguish proposed from active demand, request weighting cannot be applied consistently, or protected classes repeatedly exhaust their reservation. Because the first step is a shadow replay, rollback consists of discarding the candidate classifications and retaining the original engagement record unchanged."},"negative_tests":{"strongest_counterevidence":"Historical bursts may not predict overload: request volume could be unrelated to response delay once request complexity, owner availability, close-calendar conflicts, and evidence dependencies are considered.","problem_falsifier":"The stated problem is falsified if timestamped engagement records show that process-owner groups generally receive requests at or below their demonstrated completion capacity and that observed delays arise primarily from unclear requests, unavailable records, approval dependencies, or substantive accounting disputes rather than intake rate.","intervention_falsifier":"The intervention is falsified if a shadow replay cannot define limits that reduce above-capacity activations without delaying deadline-critical evidence, increasing projected fieldwork completion time, creating a large hidden pre-admission backlog, or shifting requests into unlogged channels.","risks":["Legitimate audit bursts may be throttled during narrow evidence windows.","Request counts may hide large differences in complexity unless weights are usable and governable.","Routine requests may be mislabeled high priority to evade the cap.","A fixed weekly reset may create a release surge or strategic timing.","The visible pre-admission register may grow while active-load measures appear healthy.","Low-volume workstreams may be starved if fairness rules reward historical volume.","Client capacity estimates may be understated to defer scrutiny or overstated to avoid escalation.","The policy could impair auditor independence if capacity governance becomes client control over audit scope.","Rate limiting may conceal a need for clearer requests, earlier planning, process redesign, or additional capacity."]},"next_evidence_step":"For one completed engagement phase and no more than three process-owner groups, extract request proposal and activation timestamps, assigned owner, due date, completion date, workstream, recorded priority, and a coarse complexity category. Reconstruct weekly arrival and completion rates, simulate a small documented range of token refills and burst allowances, and report only whether overload periods, protected-request delays, pre-admission backlog, and allocation across workstreams would have changed. Do not deploy the rule or infer an effect size from this replay.","prior_art_status":"UNSEARCHED","diversity_from_prior_proposals":"Not assessed against other proposals because runtime isolation prohibits inspecting them; this candidate is independently anchored in the admission rate of external-audit evidence requests to client accounting process owners.","revision_record":{"parent_version":null,"progress_targets_addressed":[],"conceptual_changes":[],"operational_changes":[],"evidence_changes":[],"claim_changes":[]}}