{"schema_version":1,"experiment_id":"eoa_inverse_innovation_exp09_archetype_breadth150_20260804","research_id":"eoa_inverse_innovation_exp09_light_prior_art_20260804","cell_id":"emergent_pattern_detection__accounting_auditing","search_lanes":{"direct_problem_and_intervention":{"queries":["intercompany circular journal entries network analytics adjustment loops reconciliation","intercompany discrepancy migration reversals reclassification journals across entities","graph analytics intercompany transactions reconciliation journal entries circular"],"source_ids":["SRC1","SRC2","SRC4"],"no_result_note":"No retained source described the complete proposed detector: cross-period reconstruction of circular or migrating intercompany corrections after pairwise discrepancies appear cleared."},"synonyms_and_historical_terms":{"queries":["intercompany reconciliation plug accounts suspense repeated adjustments close","intercompany out-of-balance differences delayed close survey","intercompany reconciliation recurring differences root cause corrective journals"],"source_ids":["SRC1","SRC3"],"no_result_note":null},"products_practices_and_standards":{"queries":["intercompany reconciliation software transaction matching journal entries anomaly detection","site:docs.oracle.com intercompany reconciliation journal lines differences report","site:pcaobus.org journal entries other adjustments audit standard data analytics relationships accounts"],"source_ids":["SRC2","SRC3"],"no_result_note":null},"component_combination":{"queries":["audit analytics relationship graph journal entries reversal sequence detection","accounting journal entry process mining reversals reclassifications graph","intercompany reconciliation process mining journal reversal lineage network","graph neural networks auditing accounting journal entries anomaly detection"],"source_ids":["SRC3","SRC4"],"no_result_note":"Graph-based journal analytics and journal-entry risk testing were found, but the retained sources did not combine multi-entity reconciliation links, reversal lineage, cross-period migration motifs, and controller classification."}},"sources":[{"source_id":"SRC1","title":"Managing intercompany: Navigating new challenges and tax complexities","publisher":"Deloitte","url":"https://www.deloitte.com/us/en/programs/center-for-controllership/blogs/managing-intercompany-accounting-challenges-and-tax-complexities.html","source_type":"TRADE_PROFESSIONAL","claims_supported":["Intercompany operations commonly involve organizational silos, manual workarounds, inconsistent documentation, frequent reconciliations, adjustments, and overlooked transactions.","In a Deloitte webcast poll of 4,114 participants, nearly 60% identified data quality and reconciliation or transfer-pricing documentation as their greatest intercompany-tax-management challenges.","Connected monitoring, centralized governance, analytics, dashboards, and automation are recognized approaches for improving intercompany visibility and auditability."]},{"source_id":"SRC2","title":"Intercompany Reconciliation","publisher":"Oracle","url":"https://docs.oracle.com/en/cloud/saas/financials/25d/faugl/intercompany-reconciliation.html","source_type":"FIRST_PARTY_PRODUCT","claims_supported":["Oracle's established reconciliation practice compares receivable and payable entries for a pair of provider and receiver legal entities and reports their differences.","Users can drill from period summaries through sources and journal lines to general-ledger, subledger, and originating transactions.","The documented report is pair-oriented and excludes specified balancing and clearing-company lines; it does not document cross-entity, cross-period loop or migration-motif detection."]},{"source_id":"SRC3","title":"Audit Focus: Journal Entries","publisher":"Public Company Accounting Oversight Board","url":"https://pcaobus.org/resources/staff-publications/audit-focus/audit-focus-journal-entries","source_type":"OFFICIAL_GUIDANCE","claims_supported":["PCAOB staff guidance identifies period-end adjustments, unreconciled differences, intercompany transactions, consolidating adjustments, and reclassifications as relevant considerations in journal-entry testing.","Auditors are expected to understand controls, test population completeness, select entries using professional judgment, and inspect supporting evidence.","Software audit tools may aid identification, but flags do not replace substantive testing and professional judgment."]},{"source_id":"SRC4","title":"Connecting the Dots: Graph Neural Networks for Auditing Accounting Journal Entries","publisher":"American Accounting Association","url":"https://www.efk.admin.ch/wp-content/uploads/publikationen/fachtexte/20260210_connecting-the-dots_graph-neural-networks-for-auditing-accounting-journal-entries.pdf","source_type":"PRIMARY_RESEARCH","claims_supported":["The study represents journal-entry data through account co-occurrence and transaction features, learns organization-specific graph patterns, and flags deviations from an established pattern.","Its real-world evaluation indicates that graph-based full-population analysis can identify relevant, actionable journal-entry anomalies.","The approach scores journal-entry anomalies and has interpretability limitations; it does not specifically reconstruct migrating intercompany discrepancies through reversal and replacement-journal lineage."]}],"problem_evidence":{"status":"PARTLY_SUPPORTED","finding":"The surrounding problem is visible: authoritative and professional sources document siloed intercompany work, manual workarounds, frequent adjustments, unreconciled differences, period-end risk, and consolidation complexity. However, the bounded sources do not empirically demonstrate the proposal's narrower causal phenomenon of locally reasonable corrections interacting to create persistent multi-entity circular or migrating adjustment networks.","source_ids":["SRC1","SRC2","SRC3"]},"closest_prior_art":[{"name":"Oracle Intercompany Reconciliation reports","source_ids":["SRC2"],"overlap":"Centralizes intercompany entries, compares legal-entity counterparties, identifies differences, and preserves drill-down lineage to journals and source transactions.","remaining_difference":"The documented workflow is pairwise and difference-led; it does not describe linking successive reversals, reclasses, and replacement journals across several entities and periods to detect loops after pairwise balances temporarily clear."},{"name":"Graph Neural Network Auditing framework","source_ids":["SRC4"],"overlap":"Transforms journal-entry relationships into graphs, learns organization-specific normal patterns, performs full-population anomaly detection, and presents items for auditor interpretation.","remaining_difference":"Its detection target is anomalous journal entries or graph representations, not a prespecified multi-hop motif representing migration of one unresolved intercompany amount through interacting local corrections with explicit reversal lineage."},{"name":"PCAOB journal-entry testing practice","source_ids":["SRC3"],"overlap":"Directs attention to period-end entries, reclassifications, consolidating adjustments, intercompany accounts, unreconciled differences, complete populations, supporting evidence, and human professional judgment.","remaining_difference":"It specifies risk-based audit procedures rather than an operational cross-entity motif detector or a method for discovering circular adjustment chains."}],"prior_art_disposition":"ADJACENT_PRIOR_ART","contrastive_claim_remaining":"Relative to pairwise intercompany reconciliation and journal-entry anomaly scoring, a read-only detector that explicitly links similar-value corrections, reversals, account changes, and counterparties across entities and periods can identify auditable circular or migrating adjustment chains that remain invisible after pairwise balances temporarily clear, and can provide additional bounded investigation leads without making accounting or misconduct determinations.","contrastive_claim_falsifier":"The claim fails if an existing product or published method already performs that explicit cross-entity, cross-period lineage reconstruction, or if blinded retrospective review shows that motif packets add no reproducible investigation leads beyond Oracle-style pairwise reconciliation and ordinary journal-entry anomaly selection, including after documented FX, recurring eliminations, automated reversals, and timing conventions are controlled.","gates":{"adequate_source_search":{"status":"PASS","rationale":"The bounded search covered the proposal directly, older and synonymous reconciliation language, current products and professional practice, official audit guidance, and combinations involving graph analytics, journal-entry relationships, reversals, and process lineage. Exactly four opened sources from four publishers were retained, including primary research, official guidance, and first-party documentation.","source_ids":["SRC1","SRC2","SRC3","SRC4"]},"supported_problem":{"status":"PASS","rationale":"The problem gate passes at partial support: silos, manual workarounds, frequent adjustments, reconciliation difficulty, intercompany risk, and unreconciled differences are independently visible, although circular migration itself remains unverified.","source_ids":["SRC1","SRC2","SRC3"]},"distinct_testable_claim":{"status":"PASS","rationale":"The remaining claim is narrower than the closest sources and falsifiable: test whether explicit multi-hop reversal and counterparty motifs reveal independently validated leads beyond pairwise difference reports and entry-level graph anomalies.","source_ids":["SRC2","SRC4"]},"bounded_next_test":{"status":"PASS","rationale":"A retrospective extract from three entities and two closed periods, a preregistered baseline and motif definition, at most 20 de-identified lineage packets, two independent controller reviews, and comparison with existing reconciliation outputs is bounded and directly measures incremental lead yield, reviewer agreement, lineage completeness, and false-positive explanations.","source_ids":["SRC2","SRC3","SRC4"]},"no_obvious_safety_or_authority_stop":{"status":"PASS","rationale":"The proposed first test is read-only, access-controlled, identity-suppressed, retrospective, and advisory. Accounting conclusions remain with the group controller, source evidence remains necessary, and the stated halt rules address access failure, missing lineage, and punitive use. No obvious authority or safety stop is apparent for that bounded design.","source_ids":["SRC3"]}},"screen_survival":true,"world_novelty_boundary":"This bounded public-web screen supports only an adjacent-prior-art disposition and a testable residual contrast. It cannot establish world novelty, patentability, freedom to operate, market size, expert acceptance, deployment feasibility, or realized accounting value."}