{"schema_version":1,"research_id":"eoa_inverse_innovation_exp06_external_evaluation_20260803","source_assessment_id":"predictive_residual_processing__accounting_auditing:P2:v0","cell_id":"predictive_residual_processing__accounting_auditing","search_queries":["intercompany reconciliation challenges close process survey intercompany accounting","site:docs.oracle.com intercompany reconciliation matching transactions consolidation","site:help.sap.com intercompany matching reconciliation ICMR","IAS 24 intercompany transactions eliminated consolidated financial statements official","intercompany reconciliation predicted expected transactions residual exception matching baseline accounting product","BlackLine intercompany predictive guidance matching recurring transactions","intercompany accounting survey reconciliation elimination challenges controller expressed need","academic research intercompany reconciliation automation matching transactions","site:pcaobus.org standards audit evidence completeness accuracy information produced by company","site:pcaobus.org AS 2201 information technology controls financial reporting completeness accuracy","site:bls.gov accountants auditors median pay 2025 occupational outlook","Deloitte intercompany accounting survey reconciliation elimination percentages challenges"],"sources":[{"source_id":"S1","title":"Managing intercompany: Navigating new challenges and tax complexities","publisher":"Deloitte US Center for Controllership","url":"https://www.deloitte.com/us/en/programs/center-for-controllership/blogs/managing-intercompany-accounting-challenges-and-tax-complexities.html","source_class":"AUTHORITATIVE_SECONDARY","publication_date":"n.d.","accessed_at":"2026-08-03","claims_supported":["A poll of 4,114 finance, accounting, and tax participants found that nearly 60% identified data quality and reconciliation or transfer-pricing documentation among their greatest intercompany-management challenges.","Seventy percent believed tax and controllership could collaborate more effectively, and 70% selected AI, machine learning, analytics, or dashboards as promising responses.","Deloitte identifies reconciliation, elimination, data completeness, governance, cross-functional collaboration, automation, and auditability as intercompany priorities."]},{"source_id":"S2","title":"IAS 24 Related Party Disclosures","publisher":"IFRS Foundation","url":"https://www.ifrs.org/content/dam/ifrs/publications/pdf-standards/english/2022/issued/part-a/ias-24-related-party-disclosures.pdf?bypass=on","source_class":"STANDARD","publication_date":"2022","accessed_at":"2026-08-03","claims_supported":["Intragroup related-party transactions and outstanding balances generally are eliminated when consolidated financial statements are prepared.","Related-party transactions and balances remain relevant in individual-entity financial statements, supporting preservation of entity-level accounting records rather than replacement by a group residual."]},{"source_id":"S3","title":"Intercompany Matching and Reconciliation","publisher":"SAP","url":"https://help.sap.com/docs/SAP_S4HANA_CLOUD/0fa84c9d9c634132b7c4abb9ffdd8f06/a8172f85383947bd8538bf778e771874.html?locale=en-US","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"2026-08","accessed_at":"2026-08-03","claims_supported":["SAP ICMR is an existing built-in practice intended to accelerate company-to-corporate close.","It performs real-time matching and reconciliation using configurable rules, automation, communication, and workflow.","SAP explicitly frames latency in resolving intercompany discrepancies as a problem."]},{"source_id":"S4","title":"Managing Intercompany Matching Reports","publisher":"Oracle","url":"https://docs.oracle.com/en/cloud/saas/financial-consolidation-cloud/usfcc/managing_intercompany_matching_reports.html","source_class":"OFFICIAL_PRODUCT_DOCUMENTATION","publication_date":"n.d.","accessed_at":"2026-08-03","claims_supported":["Oracle already provides intercompany matching reports for analysis and audit purposes.","The reports calculate entity-partner variances from requested database data and can suppress dimensions or data cells.","This is direct prior art for variance-focused intercompany review, but the documented report still retrieves full requested data rather than exchanging a reconstructive predicted-state residual."]},{"source_id":"S5","title":"BlackLine Announces Industry’s First AI-Enabled Intercompany Accounting Capabilities","publisher":"BlackLine","url":"https://www.blackline.com/about/press-releases/2023/blackline-announces-industrys-first-ai-enabled-intercompany-accounting-capabilities/","source_class":"COMMERCIAL_FIRST_PARTY","publication_date":"2023-09-12","accessed_at":"2026-08-03","claims_supported":["BlackLine Predictive Guidance analyzes transactional data to predict intercompany issues before booking, highlights high-risk transactions, explains risk factors, and recommends corrections.","The initially announced scope included non-trade, services-related transactions, closely overlapping the candidate's proposed first use case.","BlackLine identifies high transaction volume, cross-border, currency, ERP, data-quality, omission, delay, and manual-work problems and reported more than 4,200 platform customers."]},{"source_id":"S6","title":"AS 1105: Audit Evidence","publisher":"Public Company Accounting Oversight Board","url":"https://pcaobus.org/oversight/standards/auditing-standards/details/AS1105","source_class":"STANDARD","publication_date":"2010-08-05; amended through 2025-10-01","accessed_at":"2026-08-03","claims_supported":["Auditors must obtain sufficient appropriate evidence, with reliability affected by source, controls, and directness.","Company-produced information used as audit evidence must have accuracy and completeness tested, or relevant controls tested, and must be sufficiently precise and detailed.","A residual packet cannot safely substitute for source records or independent audit evidence without validated completeness, accuracy, provenance, and controls."]},{"source_id":"S7","title":"Finance and Insurance: NAICS 52 — Employment, Wages, and Employer Compensation Costs","publisher":"U.S. Bureau of Labor Statistics","url":"https://www.bls.gov/iag/tgs/iag52.htm","source_class":"GOVERNMENT_OR_REGULATOR","publication_date":"2026","accessed_at":"2026-08-03","claims_supported":["For 2025, accountants and auditors in finance and insurance had mean wages of $48.15 per hour and $100,150 annually.","Finance-and-insurance private-industry total compensation averaged $73.30 per hour in the first quarter of 2026.","These figures provide a labor-resource basis for broad 2026-USD estimates, but not vendor-license or organization-specific integration prices."]},{"source_id":"S8","title":"Simple Multi-Party Set Reconciliation","publisher":"arXiv; Michael Mitzenmacher and Rasmus Pagh","url":"https://arxiv.org/abs/1311.2037","source_class":"PRIMARY_RESEARCH","publication_date":"2015-04-22","accessed_at":"2026-08-03","claims_supported":["Set-reconciliation research demonstrates that parties holding large, mostly overlapping sets can communicate in proportion to their differences rather than total set size.","The paper supplies technical precedent for difference-proportional reconciliation messages.","The work does not address accounting semantics, predicted amounts, ledger authority, audit evidence, shared-model errors, or full-schedule fallback."]}],"problem_evidence":{"support":"STRONG","rationale":"The problem is visible across independent evidence: a large Deloitte practitioner poll identifies reconciliation and data quality as major intercompany challenges; SAP and BlackLine explicitly market products to reduce discrepancy latency, manual work, and close delays; IFRS requirements make correct treatment of intragroup activity consequential. Evidence supports the general problem, but not the candidate's narrower assertion that repeatedly transmitting predictable schedule content is itself a material share of the burden.","source_ids":["S1","S2","S3","S5"]},"stakeholder_evidence":{"support":"MODERATE","rationale":"Corporate controllership, entity accounting, tax, and finance-system teams are credible adopters or authorizers. Deloitte documents expressed practitioner demand for better reconciliation, collaboration, automation, and analytics, while SAP and BlackLine demonstrate purchasing and implementation activity. No named organization, controller, or entity pair has expressed willingness to pilot this specific residual-exchange protocol or release the necessary bilateral ledger data.","source_ids":["S1","S3","S5"]},"prior_art":{"proximity":"SUBSTANTIAL_COLLISION","closest_analogues":[{"name":"BlackLine Intercompany Predictive Guidance","similarity":"Predicts intercompany transaction failures before booking, prioritizes risky transactions, explains risk factors, and initially targets services-related non-trade transactions—the candidate's proposed first scope.","remaining_difference":"The public description does not establish a jointly frozen bilateral expected-state model, transmission of reconstructive signed residuals instead of declared schedules, checksum-gated interpretation, independently selected full-schedule sampling, or mandatory model-failure decompression.","source_ids":["S5"]},{"name":"SAP Intercompany Matching and Reconciliation","similarity":"Provides real-time intercompany matching, configurable reconciliation rules, discrepancy workflow, communication, automation, and earlier resolution between company close and corporate close.","remaining_difference":"The documented system matches financial data; it does not establish that counterparties communicate only model-relative residuals or reconstruct one another's declared states from a synchronized predictor with raw-channel audits.","source_ids":["S3"]},{"name":"Oracle Intercompany Matching Reports","similarity":"Calculates entity-partner variances, supports audit-oriented analysis, and permits report suppression of dimensions or cells.","remaining_difference":"It is a variance report over retrieved database data, not a bilateral predictive codec with explicit model synchronization, heartbeat semantics, independent bookings, and full-state fallback.","source_ids":["S4"]},{"name":"Difference-proportional set reconciliation","similarity":"Supplies a technical architecture in which communication scales with differences between large, mostly overlapping party-held sets rather than total population size.","remaining_difference":"It reconciles sets of keys rather than signed accounting states and supplies none of the accounting, tax, authorization, evidentiary, model-drift, or shared-error controls required here.","source_ids":["S8"]}],"distinctive_claim_remaining":"For preapproved recurring bilateral service-fee or cost-allocation flows, a checksum-synchronized expected-state-plus-residual protocol can reduce transmitted/reviewed content and bilateral reviewer time versus complete-schedule matching, while remaining noninferior on consolidation-relevant discrepancy recall because every declared state is reconstructible, missingness triggers fallback, and independently sampled complete schedules expose shared-model omissions. The claim is falsified by any material conventional finding omitted by the residual path, reconstruction outside the declared tolerance, unreliable protected-class bypass, or total governance and maintenance effort equal to or greater than the full-schedule comparator.","confidence":"MODERATE"},"implementation_evidence":{"support":"MODERATE","rationale":"Existing SAP, Oracle, and BlackLine products show that cross-entity extraction, configurable matching, variance computation, workflow, exception prioritization, and prebooking prediction are technically and operationally feasible. Difference-proportional reconciliation is also technically established. However, no source validates the complete candidate architecture on bilateral accounting data. Reliable identity mapping, reciprocal semantics, agreement and calendar versioning, missing-record detection, tax-sensitive bypasses, access control, evidence retention, and independent raw-sample governance remain implementation requirements. PCAOB standards make clear that residuals cannot replace completeness-tested source evidence.","source_ids":["S3","S4","S5","S6","S8"]},"scores":{"meaningful_impact":{"score":4,"rationale":"Intercompany reconciliation visibly affects close speed, data quality, tax documentation, and consolidation reliability; reducing repeated review could matter for high-volume groups. The candidate-specific effect size is unmeasured.","source_ids":["S1","S3","S5"]},"stakeholder_pull":{"score":3,"rationale":"Practitioners express demand for better reconciliation, collaboration, and automation, and established products demonstrate adoption. No stakeholder has requested this exact bilateral residual protocol.","source_ids":["S1","S3","S5"]},"incremental_advantage":{"score":2,"rationale":"Predictive prioritization, automated matching, variance reporting, and exception-focused workflows already exist. Only the reconstructive residual exchange plus independent full-state audit and forced fallback remains plausibly incremental, and its net advantage is untested.","source_ids":["S3","S4","S5"]},"distinctiveness_plausibility":{"score":2,"rationale":"No exact public match was found for the full control bundle, but substantial functional collision exists and the bounded search cannot establish world novelty.","source_ids":["S3","S4","S5","S8"]},"technical_implementability":{"score":4,"rationale":"Matching, prediction, workflow, and difference-proportional messaging are demonstrated components. The hard parts are data semantics, synchronization, shared-error detection, and controlled fallback rather than basic algorithmic feasibility.","source_ids":["S3","S4","S5","S8"]},"adoption_authority_feasibility":{"score":3,"rationale":"Group controllership and entity controllers can plausibly authorize a read-only shadow replay, while tax and audit functions can constrain protected classes and sampling. Cross-entity agreement and system access create material coordination burden.","source_ids":["S1","S6"]},"evidence_readiness":{"score":2,"rationale":"Complete historical schedules and booked ledgers could support a retrospective replay, but no public comparative dataset, measured workflow baseline, confirmed data partner, or tested acceptance threshold was found.","source_ids":["S3","S4","S5"]},"safety_net_benefit":{"score":4,"rationale":"Retained complete records, independent full-schedule samples, completeness markers, checksum gates, and automatic fallback directly address shared-model and missingness risks. These safeguards must be empirically tested and cannot make residuals sufficient audit evidence by themselves.","source_ids":["S6"]},"scalability":{"score":3,"rationale":"Difference-proportional communication and existing enterprise intercompany platforms support scaling in principle, but pair-specific agreements, mappings, currencies, tax rules, and approvals may make governance cost grow faster than message savings.","source_ids":["S1","S3","S5","S8"]}},"score_confidence":"MODERATE","costs":{"first_evidence":{"band_2026_usd":"10K_TO_50K","scope":"Pre-registered read-only replay of one completed close for three entity pairs, including frozen baselines, two comparators, complete-population scoring, random and risk-stratified samples, and reviewer-time measurement.","confidence":"MODERATE","assumptions":["Approximately 200-500 loaded labor hours across controllership analytics, entity accounting, data engineering, tax review, and internal audit.","Existing ledger extracts, agreements, mappings, and complete schedules are already available in a secure analytics environment.","No production integration, official counterparty confirmation, posting, or vendor license purchase is included.","The BLS finance-and-insurance total-compensation benchmark of $73.30 per hour implies roughly $14,700-$36,650 before incidental secure-compute costs."],"source_ids":["S7"]},"initial_deployment_startup":{"band_2026_usd":"250K_TO_1M","scope":"Controlled shadow prototype for a bounded recurring-flow scope: extraction connectors, common data model, baseline registry, checksum and heartbeat logic, reconstruction service, dashboards, access controls, audit logging, validation, and control documentation.","confidence":"LOW","assumptions":["Roughly two to five loaded FTE-years or equivalent contractor effort.","The estimate excludes major ERP replacement and assumes reusable identity, chart, agreement, and calendar master data.","Public sources provide no product quote or implementation benchmark; this is a resource-equivalent range anchored to BLS compensation."],"source_ids":["S3","S4","S6","S7"]},"operational_launch":{"band_2026_usd":"1M_TO_5M","scope":"Production launch across an initial multi-entity cohort, including hardened integrations, segregation of duties, bilateral approvals, tax-sensitive scope controls, fallback capacity, parallel closes, security and control testing, training, incident response, and audit support.","confidence":"LOW","assumptions":["Approximately eight to twenty loaded FTE-years across finance, engineering, security, tax, audit, change management, and program governance, plus software and infrastructure.","At least two parallel close cycles are required before any operational suppression.","Heterogeneous ERPs and local statutory requirements can move the cost within or above this band.","No vendor pricing or organization-specific systems inventory was publicly available."],"source_ids":["S1","S3","S5","S6","S7"]},"annual_recurring":{"band_2026_usd":"250K_TO_1M","scope":"Ongoing model and mapping governance, agreement refresh, monitoring, full-schedule sampling, fallback exercises, incident handling, access reviews, control testing, tax review, and user support for the initial cohort.","confidence":"LOW","assumptions":["Approximately two to five loaded FTE equivalents annually, with additional burst capacity during close and fallback periods.","Complete source-record retention and independent audit sampling continue; they are not replaced by residual storage.","License, cloud, and audit fees are organization-specific and may push a large deployment above the band."],"source_ids":["S6","S7"]}},"verified_pipeline_gates":{"externally_supported_problem":{"status":"YES","reason":"Independent practitioner evidence and multiple established products support intercompany reconciliation, discrepancy latency, data-quality, manual-work, and close-delay problems, although predictable-content burden has not been isolated.","source_ids":["S1","S3","S5"]},"externally_credible_adopter_or_authorizer":{"status":"YES","reason":"Group controllership, legal-entity controllers, and tax and finance-system leaders are credible authorizers or adopters; Deloitte documents their expressed need and SAP and BlackLine demonstrate an active adoption category. This does not constitute a pilot commitment.","source_ids":["S1","S3","S5"]},"distinct_testable_incremental_claim":{"status":"YES","reason":"The remaining claim specifies a bounded recurring-flow scope, full-schedule and simple-baseline comparators, measurable communication and reviewer-time savings, noninferior discrepancy recall, reconstruction fidelity, and explicit falsifiers.","source_ids":["S3","S4","S5","S8"]},"bounded_next_evidence_step":{"status":"YES","reason":"A single completed-close replay over three preselected entity pairs can be pre-registered, run read-only, compared against complete schedules, and terminated without affecting books or confirmations.","source_ids":["S6"]},"no_unresolved_safety_or_authority_stop":{"status":"YES","reason":"The authorized first step is read-only and retains complete schedules, entity booking authority, source provenance, tax review, independent sampling, and auditor access. PCAOB requirements prevent treating residual packets as replacement audit evidence. Production suppression remains unauthorized pending validation.","source_ids":["S2","S6"]},"credible_cost_scope_and_range":{"status":"YES","reason":"All four ranges state included scope and labor assumptions and are anchored to a 2026 government compensation benchmark. Confidence is low beyond the replay because vendor pricing, system heterogeneity, and local-control requirements remain unknown.","source_ids":["S7"]}},"next_evidence_step":"With a participating group controller, pre-register a read-only retrospective replay of one completed close for three entity pairs and only recurring service-fee or cost-allocation flows. Freeze each prediction using information timestamped before that close. Compare (A) the existing complete-schedule process, (B) simple prior-period or rules-based variance matching representative of established practice, and (C) synchronized baseline-plus-residual reconstruction. Score the complete population for material-discrepancy recall, precision, reconstruction error, unmatched and missing items, protected-class bypass, version and heartbeat failures, fallback frequency, bytes or rows presented, bilateral reviewer minutes, consolidation handoffs, maintenance time, and total resource-equivalent cost. Independently draw one random transaction sample and one tax/risk-stratified sample without residual scores. Falsify the intervention if C misses any material finding found by A, exceeds the predeclared reconstruction tolerance, fails to bypass a protected item, cannot distinguish missing data from agreement, or does not reduce total reviewer-plus-maintenance effort by at least 20% relative to A while remaining noninferior on recall. The replay authorizes no posting, clearing, netting, elimination, settlement, official confirmation, or source-record suppression.","blocking_evidence":["No organization or group controller has committed proprietary bilateral ledger data, agreements, complete schedules, reviewer time, or staff for the replay.","No comparative evidence shows that predictable schedule review is a material share of current intercompany effort or that the protocol reduces total effort after model, synchronization, audit, and fallback work.","Reconstruction fidelity, material-discrepancy recall, missing-record detection, shared-baseline-error detection, and protected-class bypass have not been tested on real accounting populations.","No organization-specific systems inventory, vendor quote, control-design assessment, data-residency review, transfer-pricing review, or local statutory analysis supports the deployment and recurring cost estimates.","The public BlackLine description creates substantial prior-art overlap; product-level technical documentation or a direct feature comparison is needed to determine whether the remaining architecture is operationally distinct."],"research_disposition":"PARTNERED_RESEARCH_PROGRAM","world_novelty_boundary":"A bounded web search found established real-time intercompany matching, variance reporting, exception workflows, prebooking predictive guidance, and generic difference-proportional reconciliation. It did not verify an exact implementation combining independently booked bilateral states, a frozen shared predictor, reconstructive signed residual messages, checksum and heartbeat gates, independent full-schedule sampling, and mandatory fallback. Absence from these eight sources is not evidence of world novelty. Patentability, freedom to operate, market size, realized impact, and comprehensive product, patent, standards, and non-English literature coverage remain unmeasured.","arm":"COMPLETE_PROPOSAL_PORTFOLIO","candidate_version":0,"controller_recommendation":{"action":"STOP_EMPIRICAL_RESEARCH_NEEDED","repairable":false,"material_progress_observed":true,"progress_targets":["Secure written participation from one group controller, three entity-pair controllers, tax, systems administration, and internal audit for a read-only completed-close replay.","Pre-register materiality, reconstruction tolerance, noninferiority margin, 20% net-effort threshold, sampling design, protected classes, and automatic falsifiers before viewing outcomes.","Measure complete-population discrepancy recall and total reviewer, maintenance, audit, and fallback effort against both complete-schedule and simple rules-based comparators.","Obtain product-level documentation or vendor clarification on BlackLine Predictive Guidance and SAP ICMR to determine whether synchronized reconstructive residual exchange and independent raw-state fallback are genuinely incremental.","Produce organization-specific production cost, data-governance, cybersecurity, transfer-pricing, statutory-record, and audit-control assessments before any operational suppression."],"reason":"Open-book research supports the problem, credible authorizers, component feasibility, safeguards, and a bounded test, but cannot establish the candidate's incremental benefit or safety. The decisive evidence requires proprietary bilateral records, human workflow measurement, controlled replay, and product-level feature verification. Under the required controller rule, that fieldwork dependency requires STOP_EMPIRICAL_RESEARCH_NEEDED, and every STOP is non-repairable in this evaluation cycle."},"proposal_index":2}