AI-assisted corporate compensation settlement optimization system

Abstract

This system provides a compensation settlement optimization system that determines fair compensation by integrating execution results, quality evaluations, responsibility boundaries, and compensation conditions after AI agents or AI processing modules belonging to multiple legal entities have performed collaborative processing. [Solution] The AI-assisted inter-company compensation settlement optimization system comprises a performance acquisition unit 361, a settlement condition acquisition unit 362, a settlement candidate generation unit 363, a settlement evaluation unit 364, a settlement plan determination unit 365, a settlement execution control unit 366, and a recording unit 367. Based on execution performance data and settlement conditions, it generates multiple compensation settlement candidates, calculates evaluation values ​​based on the degree of conformity to contract conditions, the degree of consistency between income and expenditure, the degree of contribution reflection, the appropriateness of reduction, the appropriateness of reservations, or the degree of dispute prevention, and finalizes the compensation settlement plan. The recording unit stores the candidates, evaluation values, selection reasons, and settlement plans as audit records with hash value chains.

Description

Technical Field

[0001] The present invention relates to a technology for supporting corporate operations using artificial intelligence (AI). In particular, for cases or processing units executed in cooperation by AI agents or AI processing modules belonging to multiple corporations, based on execution results, resource usage records, result acceptance results, responsibility demarcation information, and consideration conditions, multiple consideration settlement candidates are generated, compared and evaluated, and then an AI-supported inter-corporate consideration settlement optimization system that determines a consideration settlement plan is concerned.

Background Art

[0002] In recent years, in corporate operations such as contract review, procurement, financial management, registration, translation, audit assistance, and others, a configuration in which AI is responsible for case reception, document analysis, processing execution, review support, and audit record generation has been used.

[0003] On the other hand, in conventional inter-corporate settlements, it is often the case that the fixed amount determined at the time of the contract is paid as it is, or manual adjustment is made through additional agreements. However, with such a configuration, a mechanism for mechanically generating multiple settlement candidates based on execution records (resource usage amount, result item quality, deadline compliance rate, number of re-executions, responsibility switching history, etc.) and comparing and selecting delivery results, quality evaluation, SLA violation, re-execution burden, responsibility demarcation, etc. as evaluation indicators is not sufficient.

[0004] However, after inter-corporate collaborative processing is executed, a mechanism for finally determining as a whole how much to pay to whom, where to reduce the amount, and where to withhold, taking into account the actual resource usage amount, the quality of the result, the deadline fulfillment status, the presence or absence of re-execution, the change in responsibility demarcation, etc. is not sufficient.

[0005] How to settle the final compensation after collaborative processing largely depends on each entity's contribution, degree of responsibility, degree of outcome fulfillment, degree of resource excess, burden of re-execution, and reserve conditions. A fixed-amount settlement is appropriate if all processing proceeds as planned, but it is inappropriate to pay a fixed amount if some deliverables are not fulfilled, if some entities require re-execution due to resource excess, or if there are remaining cases with ambiguous responsibilities.

[0006] For example, if corporations A, B, C, and D perform a collaborative process, and corporation C triggers a re-execution due to resource overload, resulting in an additional burden on corporation D, or if some deliverables are delivered incompletely, a system where each corporation is simply paid a fixed amount as agreed upon in the contract would not reflect their contribution, degree of responsibility, and any reserved conditions, potentially leading to unfair settlements or disputes.

[0007] Therefore, there is a need for technology that, after coordinating processing between corporations, acquires delivery results, execution results, quality evaluations, responsibility transfer history, number of re-executions, etc., generates multiple settlement options (full settlement options, reduced settlement options, retained settlement options, resubmission request options, etc.), compares and evaluates them based on factors such as the degree of conformity to contract terms, the degree of balance between income and expenses, the degree of reflection of contributions, the degree of dispute prevention, and the degree of clarity of responsibility division, in order to determine an appropriate settlement plan. [Prior art documents] [Patent Documents]

[0008] [Patent Document 1] Japanese Patent Application No. 2026-005527 discloses a corporate management system equipped with machine-readable articles of incorporation, an AI execution engine, audit nodes, and an information sharing platform. While it shows the overall architecture of corporate management, it does not disclose a consideration settlement optimization system that integrates execution results, quality evaluation, and responsibility demarcation, generates and evaluates multiple settlement candidates, and determines, executes, and records settlement plans.

[0009] [Patent Document 2] Japanese Patent Application No. 2026-007023 discloses data items such as AI_registry_id, task_id, timestamp, and evidence_ref, as well as task generation and evidence recording. While it provides the basics of AI agent management, it does not disclose the acquisition of execution results across multiple legal entities, normalization of settlement conditions, generation and evaluation of settlement candidates, determination of settlement plans, and integrated execution control and recording for consideration settlement optimization.

[0010] [Patent Document 3] Japanese Patent Application No. 2026-047022, "Prior Consent Gate System," discloses functions for verifying prior consent conditions, confirming the integrity of supporting documents, and assisting audits. While it provides the basics of consent management and document recording, it does not disclose the normalization of settlement conditions that integrates execution results, quality evaluation, and responsibility boundaries, the generation and evaluation of multiple settlement candidates, or the determination, execution control, and recording of settlement plans.

[0011] [Patent Document 4] U.S. Patent Application Publication No. 2022 / 0318752 (US20220318752A1) "Systems and Methods for Real-Time Contract Settlement" discloses systems and methods for real-time contract settlement. It shows a mechanism for verifying, pricing, and settling medical insurance claims in real time using smart contracts and distributed ledgers, but it does not describe the integration of execution results, quality evaluation, and liability boundaries between corporations, normalization of settlement conditions, generation and evaluation of multiple settlement candidates, settlement plan determination, and execution control and recording.

[0012] [Patent Document 5] U.S. Patent No. 11410233 (US11410233B2) "Blockchain Technology to Settle Transactions" discloses real-time settlement of securities transactions using blockchain technology. It outlines the basics of immediate gross settlement of securities trades using cryptographically signed clearing instructions and a hierarchical distributed ledger, but it does not disclose consideration settlement optimization that integrates the acquisition of execution results between corporations, normalization of settlement conditions, generation and evaluation of multiple settlement candidates, settlement plan determination, and execution control and recording. [Non-patent literature]

[0013] [Non-Patent Document 1] S. Nakamoto, "Bitcoin: A Peer-to-Peer Electronic Cash System", 2008 https: / / bitcoin.org / bitcoin.pdf Bitcoin (P2P electronic cash system). While it discusses settlement using a distributed ledger, it does not address the normalization of settlement conditions by integrating execution results, quality evaluation, and responsibility boundaries between corporations, nor does it deal with the generation and evaluation of multiple settlement candidates or the determination of settlement plans.

[0014] [Non-Patent Document 2] M. Swan, "Blockchain: Blueprint for a New Economy", O'Reilly, 2015 https: / / www.oreilly.com / library / view / blockchain-blueprint-for / 9781491920480 / This book comprehensively discusses blockchain applications, but does not cover the implementation of execution performance, quality evaluation, and responsibility demarcation integration between corporations, settlement condition normalization, generation and evaluation of multiple settlement candidates, settlement plan determination, and execution control and recording. [Overview of the Initiative] [Problems that the invention aims to solve]

[0015] The first problem of the present invention is to obtain execution results, resource usage records, result acceptance results, deadline fulfillment results, and responsibility demarcation information in a common format that can be used for settlement judgment.

[0016] The second problem of the present invention is not to directly output a single settlement result, but to generate a plurality of countervalue settlement candidates including a fixed amount allocation candidate, a contribution ratio allocation candidate, a quality correction candidate, a responsibility apportionment candidate, or a retention candidate to make them comparable.

[0017] The third problem of the present invention is to enable recalculation of all or part of the countervalue settlement plan when there is a return of deliverables, re-execution, deadline delay, resource excess, or responsibility demarcation change.

[0018] The fourth problem of the present invention is to save the candidate generation process, evaluation values, selection reasons, and the determined countervalue settlement plan as audit records for later verification.

[0019] The present invention has been made to solve the above problems, and for projects or processing units executed in cooperation by AI agents or AI processing modules belonging to multiple corporations, based on execution records, result acceptance results, responsibility demarcation, and countervalue conditions, it aims to provide a mechanism for determining a highly valid countervalue settlement plan.

Means for Solving the Problems

[0020] The AI-assisted inter-corporate countervalue settlement optimization system according to one aspect of the present invention includes a performance acquisition unit, a settlement condition acquisition unit, a settlement candidate generation unit, a settlement evaluation unit, a settlement plan determination unit, a settlement execution control unit, and a recording unit.

[0021] The performance acquisition unit can obtain execution performance data including a project identifier, the responsible corporation, the processing result, the resource usage record, the result acceptance result, the deadline fulfillment result, or re-execution occurrence information.

[0022] The settlement conditions acquisition unit can acquire at least a portion of the consideration conditions, reduction conditions, retention conditions, incentive conditions, responsibility demarcation conditions, or settlement deadline conditions set for each of the multiple legal entities.

[0023] The settlement candidate generation unit can generate a plurality of settlement candidates based on the execution performance data and the conditions obtained by the settlement condition acquisition unit.

[0024] The aforementioned multiple compensation settlement candidates may include fixed-amount allocation candidates, contribution-proportional allocation candidates, quality-results-adjusted allocation candidates, burden-sharing candidates according to responsibility boundaries, or candidates for retaining a portion of the compensation.

[0025] The settlement evaluation unit can calculate an evaluation value for the plurality of settlement candidates based on at least one of the following: degree of conformity to contract terms, degree of consistency between income and expenditure, degree of reflection of contribution, degree of appropriateness of reduction, degree of appropriateness of reservation, or degree of suppression of dispute occurrence.

[0026] The settlement plan determination unit can select one of the multiple settlement candidates based on the evaluation value and finalize the settlement plan.

[0027] The settlement execution control unit may generate at least one of the following based on the settlement plan: billing data, allocation data, payment withholding data, settlement confirmation notice data, or re-settlement request data.

[0028] The recording unit stores execution data, settlement conditions, settlement candidates, evaluation values, selection reasons, consideration settlement plans, recalculation requests, and settlement results as audit records, and may add hash_value, policy_version, or time information as needed.

[0029] The aforementioned compensation conditions, reduction conditions, reservation conditions, or incentive conditions may include fixed compensation amounts, usage-based unit rates, conditions fixed after acceptance of results, reduction rates for failure to meet SLAs, burden ratios for re-execution, minimum guaranteed amounts, success fee rates, or whether reservations are required in the event of a dispute.

[0030] The aforementioned liability demarcation conditions may include the scope of responsibility for each requesting entity, processing entity, sub-delegated entity, review entity, and audit entity, classification of the cause of the anomaly, exemption conditions, burden limits, or recalculation priority.

[0031] The settlement candidate generation unit may generate a group of settlement candidates, each assigned a candidate_id for the same case, and associate the allocation amount, reduction amount, retained amount, confirmation conditions, or release conditions with each candidate. For example, it may generate a fixed amount candidate that applies the fixed amount stipulated in the contract as is, a proportional candidate that adjusts the allocation amount according to the contribution_ratio of each corporation, a quality correction candidate that reduces the amount in part according to the outcome acceptance score or the result of failing to meet the SLA, a responsibility allocation candidate that apportions the burden according to the responsibility division for the cause of re-execution, and a retained candidate that retains a portion of the consideration until the audit is completed or the dispute is resolved. Furthermore, each candidate may be associated with and stored with settlement_type, distribution_result, deduction_result, hold_result, hold_release_condition, responsibility_scope_ref, and evidence_ref.

[0032] The settlement evaluation unit may reduce the evaluation score or exclude candidates for which it has detected a high likelihood of breaching contract conditions, exceeding the budget, mismatching the conditions for accepting results, inconsistency with the division of responsibilities, or a high likelihood of disputes. In this case, the settlement evaluation unit may maintain a breakdown of the degree of conformity to contract conditions, consistency of income and expenditure, degree of contribution reflection, justification of reduction, justification of reservations, and degree of dispute prevention as score_breakdown, and may store the basis for deduction or exclusion in an auditable form by assigning reason_code, deduction_basis_ref, liability_cap_ref, hold_release_condition, dispute_hold_flag, or evidence_ref. For example, reason_codes such as SLA_MISS_DEDUCTION_REQUIRED, UNACCEPTED_OUTPUT_HOLD_REQUIRED, LIABILITY_SCOPE_MISMATCH, BUDGET_OVER_TOTAL_CAP, or DISPUTE_RISK_HIGH may be used.

[0033] The settlement plan determination unit may apply different settlement evaluation criteria or weighting policies depending on the type of case, confidentiality classification, the requesting corporation's priority policy, the recipient corporation's scope of responsibility, or the terms of continuing transactions. Furthermore, the emphasis axes used when comparing candidates may be switched based on priority_policy_id, liability_profile_id, dispute_hold_flag, continuity_preference_flag, or confidentiality_level.

[0034] The settlement execution control unit may issue a settlement recalculation request, specifying all or part of the settlement plan as the subject of recalculation, if it detects at least one of the following: return of deliverables, occurrence of re-execution, deadline delay, resource overrun, approval expiration, or change in responsibility boundaries.

[0035] The recording unit may generate a hash_value corresponding to each settlement candidate or settlement plan and save it in association with the hash_value corresponding to the preceding record.

[0036] The settlement conditions acquisition unit or settlement candidate generation unit may normalize and utilize contract terms ledgers, consideration ledgers, settlement policy ledgers, responsibility boundary ledgers, or audit ledgers managed for each of the multiple legal entities based on policy_version, item correspondence information, or schema correspondence information.

[0037] The aforementioned system may be linked with the inter-corporate condition matching and agreement generation protocol of derivative D and the resource allocation plan determination of derivative E. For the group of cases deemed permissible to execute by derivative D, derivative F may be configured to determine which corporation will receive which consideration and under what conditions, based on the execution results allocated by derivative E. [Effects of the Invention]

[0038] According to the present invention, after cooperative processing between corporations, the execution results (resource usage, deliverable quality, adherence to deadlines, number of re-executions, responsibility transfer history) are integrated, and multiple settlement candidates (full settlement candidates, reduced settlement candidates, retained settlement candidates, resubmission request candidates, etc.) are generated, compared, and evaluated. A settlement plan can then be determined based on the degree of suitability to contract conditions, balance of income and expenditure, degree of reflection of contribution, degree of dispute suppression, and degree of clarity of responsibility division. This makes it possible to allocate compensation in a manner that is more in line with reality compared to fixed-amount settlement or manual adjustment.

[0039] Furthermore, according to the present invention, since multiple settlement candidates can be compared and selected based on evaluation indicators, while simultaneously considering delivery and acceptance conditions, SLA, number of re-executions, responsibility boundaries, and reservation conditions, it is possible to deter settlements that induce overpayments, underpayments, or disputes.

[0040] Furthermore, according to the present invention, for each settlement candidate, it is possible to evaluate the degree of conformity to contract conditions (degree of outcome fulfillment, degree of adherence to deadlines, degree of SLA fulfillment), the degree of balance of income and expenditure (degree of budget conformity, prevention of overpayment and underpayment), the degree of reflection of contribution (degree of contribution of each corporation, division of responsibility), the degree of dispute suppression (degree of conformity to reserved conditions, degree of clarity of grounds for reduction), and the degree of clarity of responsibility division (history of responsibility transfer, burden of re-execution). Therefore, instead of simply paying a fixed amount at the time of the contract, it is possible to select a settlement plan that reflects the actual execution results, quality evaluation, and division of responsibility.

[0041] Furthermore, according to the present invention, a re-settlement request can be issued and recalculated if results are resubmitted or conditions change, so settlement can be applied even if changes occur in the middle of cooperative processing.

[0042] In addition, according to the present invention, the process of generating settlement candidates, evaluation values ​​(breakdown of contract condition compliance, balance of income and expenditure, degree of contribution reflection, degree of dispute suppression, and degree of clarity of responsibility division), and the reasons for the final decision are retained as audit records, making it easier to explain after the fact why a particular allocation amount or reserve amount was selected. [Brief explanation of the drawing]

[0043] [Figure 1] This is a block diagram showing the overall configuration of an AI-assisted inter-company payment settlement optimization system according to one embodiment of the present invention. [Figure 2] This is a functional block diagram showing the functional coordination between the performance acquisition unit, settlement condition acquisition unit, settlement candidate generation unit, settlement evaluation unit, settlement plan determination unit, settlement execution control unit, and recording unit. [Figure 3] This is a process flow diagram for normalizing the contract terms ledger, consideration ledger, responsibility boundary ledger, and audit ledger. [Figure 4] This is a flowchart for generating settlement candidate options, which includes fixed amount candidates, contribution-proportional candidates, quality adjustment candidates, responsibility allocation candidates, and reservation candidates. [Figure 5] This diagram shows the calculation of evaluation values ​​for each settlement candidate and the selection process for the settlement plan. [Figure 6]This flowchart illustrates the process of issuing a settlement recalculation request in response to the return of deliverables, re-execution, or changes in the division of responsibility. [Figure 7] This diagram shows an audit record structure that records settlement candidates and settlement plans with hash_value chaining. [Figure 8] This is a use case diagram illustrating an example where company A and company D collaborate on a contract review project and then settle the payment. [Modes for carrying out the invention]

[0044] 1. Overall structure As shown in Figure 1, the AI-assisted inter-corporate payment optimization system 300 of this embodiment includes corporate system A 310, corporate system B 320, corporate system C 330, corporate system D 340, and an information sharing platform 350.

[0045] Each corporate system 310, 320, 330, and 340 may have an AI execution engine, a performance notification module, a settlement condition registration module, an audit log module, and, if necessary, a manual review terminal or a legal review terminal.

[0046] The information sharing platform 350 may include an execution record ledger 351, a consideration conditions ledger 352, a responsibility boundary ledger 353, a settlement candidate record unit 354, a consideration settlement plan record unit 355, an audit record ledger 356, and a normalization rule storage unit 357.

[0047] The execution record ledger 351 may contain the following entries: task_id, entity_id, resource_usage_value, acceptance_result, deadline_result, rework_flag, evidence_ref, and timestamp.

[0048] The consideration conditions ledger 352 may store policy_id, policy_version, fixed_fee_rule, usage_fee_rule, sla_penalty_rule, hold_rule, success_bonus_rule, and settlement_deadline.

[0049] The liability demarcation ledger 353 may register requester_entity_id, provider_entity_id, review_entity_id, audit_entity_id, responsibility_scope, liability_cap, rollback_trigger, and dispute_hold_flag. The liability demarcation ledger 353 may also hold liability_profile_id for each liability type or continuity_preference_flag for continuing transactions.

[0050] The settlement candidate recording unit 354 may record candidate_id, task_id, settlement_type, distribution_result, deduction_result, hold_result, reason_code, and score_breakdown.

[0051] The settlement plan recording unit 355 may record selected_candidate_id, confirmed_distribution, confirmed_deduction, confirmed_hold_amount, release_condition, and settlement_status.

[0052] 2. Performance Acquisition Department As shown in Figure 2, the performance acquisition unit 361 receives the execution results of derivatives D and E and generates execution performance data including the case identifier, responsible company, processing result, resource usage record, result acceptance result, deadline fulfillment result, whether or not the execution was re-executed, and evidence reference.

[0053] The performance acquisition unit 361 may, if necessary, extract performance data from execution logs or audit logs held by each corporation and normalize it into a common schema.

[0054] Furthermore, if multiple corporations are involved in the same project, the performance acquisition unit 361 may calculate the contribution ratio by process or by role and use it to generate settlement candidates in a later stage.

[0055] 3. Settlement Condition Acquisition Unit The settlement conditions acquisition unit 362 acquires the consideration conditions, reduction conditions, retention conditions, incentive conditions, responsibility demarcation conditions, or settlement deadline conditions set for each of the multiple legal entities.

[0056] If the ledger item names differ among multiple legal entities, the settlement condition acquisition unit 362 may refer to the normalization rule storage unit 357 and convert, for example, "penalty_rate", "sla_deduction_rate", and "late payment reduction rate" into a common item deduction_rate (see Figure 3).

[0057] Furthermore, the settlement conditions acquisition unit 362 may select the applicable policy_version according to the effective date, and prepare the system to be comparable even if there are remaining cases based on older contract versions or old settlement rules.

[0058] 4. Settlement candidate generation unit The settlement candidate generation unit 363 generates multiple settlement candidates based on the execution performance data and settlement conditions (see Figure 4).

[0059] As an example, the settlement candidate generation unit 363 may generate (i) a fixed amount candidate that applies the fixed amount stipulated in the contract as is, (ii) a proportional candidate that allocates the amount proportionally to Corporation B and Corporation C according to the contribution_ratio, (iii) a quality adjustment candidate that reduces the amount in part according to the outcome acceptance score, SLA achievement status, or number of reworks, (iv) a responsibility allocation candidate that allocates the burden according to the responsibility division for the cause of re-execution, and (v) a reserve candidate that reserves a portion of the consideration until the dispute is resolved, the audit is completed, or the acceptance of the results is confirmed.

[0060] The settlement candidate generation unit 363 may assign a candidate_id to each candidate and store associated information such as the allocation amount, reduction amount, reserve amount, confirmation conditions, reservation release conditions, and related evidence references. Furthermore, it may store settlement_type, distribution_result, deduction_result, hold_result, hold_release_condition, responsibility_scope_ref, and evidence_ref for each candidate and use them for subsequent comparative evaluation.

[0061] Furthermore, the settlement candidate generation unit 363 may change the set of candidates according to the case type, confidentiality classification, continuing transaction terms, or the requesting corporation's priority policy.

[0062] 5. Settlement and Evaluation Department The settlement evaluation unit 364 calculates an evaluation value for each settlement candidate based on the degree of conformity to contract terms, the degree of consistency between income and expenditure, the degree of reflection of contribution, the appropriateness of reduction, the appropriateness of reservation, or the degree of suppression of dispute occurrence (see Figure 5).

[0063] For example, the degree of contractual compliance may be calculated based on the extent to which the candidate conforms to fixed-amount clauses, SLA reduction clauses, re-execution burden clauses, or reservation conditions.

[0064] The degree of balance between income and expenditure may be calculated based on the total amount paid to each of the candidate corporations, the total amount reduced, the total amount retained, and the budget utilization rate.

[0065] The degree of contribution may be calculated based on each organization's workload by process, contribution to outcome acceptance, contribution to review, or contribution to troubleshooting.

[0066] The settlement evaluation unit 364 may reduce the evaluation score or exclude candidates in which it has detected a high likelihood of breach of contract terms, budget overrun, inconsistency with the conditions for accepting results, inconsistency with the division of responsibilities, or a high likelihood of disputes.

[0067] At this time, the settlement evaluation unit 364 may maintain a breakdown of the degree of conformity to contract terms, the degree of consistency between income and expenditure, the degree of contribution reflection, the appropriateness of reductions, the appropriateness of reservations, and the degree of dispute prevention as score_breakdown, and may assign reason_code, deduction_basis_ref, liability_cap_ref, hold_release_condition, dispute_hold_flag, or evidence_ref.

[0068] 6. Settlement Planning Department The settlement plan determination unit 365 selects one of several settlement candidates based on the evaluation value and finalizes the settlement plan.

[0069] The settlement planning unit 365 may apply different settlement evaluation criteria or weighting policies depending on the type of case, confidentiality classification, the requesting entity's priority policy, the recipient entity's scope of responsibility, or the terms of continuing transactions.

[0070] For example, in ongoing transactions, the degree to which disputes arise and the maintenance of relationships are relatively emphasized; in urgent cases, the contribution to meeting deadlines and the contribution to re-implementation are emphasized; in highly confidential cases, internal self-sufficiency and clarity of responsibility are emphasized; in quality-related cases, the appropriateness of quality corrections and consistency of evidence are emphasized; and in dispute-prone cases, the appropriateness of reservations and consistency of responsibility are emphasized.

[0071] 7. Settlement Execution Control Unit The settlement execution control unit 366 generates billing data, allocation data, payment withholding data, offsetting notice, or settlement confirmation notice based on the finalized settlement plan.

[0072] The settlement execution control unit 366 may issue a settlement recalculation request to recalculate all or part of the settlement plan if it detects at least one of the following: return of deliverables, occurrence of re-execution, deadline delay, resource overrun, approval expiration, or change in responsibility boundaries (see Figure 6).

[0073] A settlement recalculation request may include the task_id to be recalculated, the affected candidate_id, the amount of impact, the reason for the change, whether a provisional reservation is necessary, and the deadline for reassessment.

[0074] The settlement execution control unit 366 may perform a partial re-settlement after issuing a re-settlement request, retaining only the finalized portion and reserving only the disputed portion.

[0075] 8. Records Section The recording unit 367 stores the execution performance data, acquired settlement conditions, generated settlement candidates, evaluation values ​​for each candidate, reasons for selection, finalized settlement plan, settlement recalculation request, and settlement results as audit records.

[0076] Each record may include task_id, candidate_id, selected_candidate_id, policy_version, entity_id, score_breakdown, reason_code, timestamp, evidence_ref, and hash_value.

[0077] The recording unit 367 may generate a hash_value corresponding to each settlement candidate or settlement plan and save it in association with the hash_value corresponding to the preceding record, thereby forming a chain structure that allows for tamper detection (see Figure 7).

[0078] Furthermore, audit records may be searched and reorganized and displayed by project, legal entity, candidate_id, responsibility boundary, or reservation status. [Examples]

[0079] Payment settlement for contract review projects As shown in Figure 8, it is assumed that the project was carried out collaboratively by derivatives D and E, with corporation A acting as the client, corporation B as the text review officer, corporation C as the clause extraction officer, and corporation D as the final audit officer.

[0080] The performance acquisition unit 361 acquires the completion of the main text review for Corporation B, the completion of clause extraction for Corporation C, the completion of the audit for Corporation D, GPU usage records, deadline fulfillment results, and results of deliverable acceptance.

[0081] The settlement candidate generation unit 363 generates (a) a fixed amount candidate to pay 120,000 yen to corporation B, 40,000 yen to corporation C, and 20,000 yen to corporation D; (b) a proportional allocation candidate to pay 110,000 yen to corporation B, 50,000 yen to corporation C, and 20,000 yen to corporation D according to contribution_ratio; (c) a quality adjustment candidate that reduces the amount in part to reflect the expected failure of corporation C to meet the SLA; (d) a responsibility allocation candidate that increases the burden on corporation C if the cause of re-execution lies with corporation C; and (e) a reserve candidate that reserves a portion of the consideration until the audit is completed.

[0082] The settlement evaluation unit 364 calculates the degree of contract condition compliance, revenue and expenditure consistency, contribution reflection, justification of reduction, justification of reservation, and degree of dispute prevention for each candidate, and stores the breakdown as score_breakdown. The settlement plan determination unit 365 switches the emphasis axis according to priority_policy_id, liability_profile_id, dispute_hold_flag, or continuity_preference_flag, and selects the candidate with the highest evaluation value.

[0083] The settlement execution control unit 366 generates billing data and allocation data based on the selected settlement plan, and the recording unit 367 saves the candidate_id, policy_version, and hash_value as audit records. [Examples]

[0084] Reduction and withholding in case of failure to meet SLA If the quality score of the clause extraction results for corporation C falls below a predetermined threshold and re-execution becomes necessary, the settlement execution control unit 366 detects the need for re-execution.

[0085] The settlement candidate generation unit 363 generates (a) a candidate to reduce the consideration for corporation C by 20%, (b) a candidate to temporarily reserve 30% of the consideration for corporation C and release it after the re-execution is completed, (c) a candidate in which corporation C partially bears the cost of the re-execution while adding the additional audit fee for corporation D, and (d) a candidate to reserve only the portion to which the acceptance of results has not been completed as the disputed part. Each candidate may have a hold_release_condition set to at least one of the following: completion of re-execution, completion of re-acceptance, completion of audit, or resolution of dispute.

[0086] The settlement evaluation unit 364 evaluates each candidate based on the contract terms, quality results, re-execution liability, and continuing transaction conditions, and selects the candidate that best balances dispute prevention and compliance. At this time, a reason code such as SLA_MISS_DEDUCTION_REQUIRED, UNACCEPTED_OUTPUT_HOLD_REQUIRED, LIABILITY_SCOPE_MISMATCH, or DISPUTE_RISK_HIGH may be attached, and the basis for deduction or exclusion may be saved.

[0087] The settlement execution control unit 366 reserves a portion of the payment according to the selected candidate and sets the conditions for releasing the reservation as completion of re-execution and completion of re-acceptance. [Examples]

[0088] Re-settling when the division of responsibility changes If a subsequent audit reveals that some of the defects initially attributed to Corporation C were actually caused by insufficient pre-processing by Corporation B, then the liability demarcation ledger 353 will be updated.

[0089] The settlement execution control unit 366 detects a change in the division of responsibility and issues a settlement recalculation request that recalculates the portion of the existing settlement plan that relates to the reduction amount and the burden ratio.

[0090] The settlement candidate generation unit 363 generates new responsibility allocation candidates based on the updated responsibility boundaries, and the settlement evaluation unit 364 re-evaluates them based on the differences from the old candidates and the consistency of the contract conditions.

[0091] The settlement plan determination unit 365 selects a new responsibility allocation candidate, and the recording unit 367 saves the difference from the old plan, the reason for the change, and the correspondence between the old and new hash_values. [Industrial applicability]

[0092] This invention is applicable to various industrial fields where AI-assisted task sharing is performed among multiple entities, including legal and contract review, accounting and tax support, registration and notification, purchasing and procurement, BPO, translation, audit assistance, collaborative research management, supply chain coordination, and more. [Explanation of Symbols]

[0093] 300 AI-assisted inter-company payment settlement optimization system 310 Corporate System A 320 Corporate B System 330 Corporate C System 340 Corporate D System 350 Information sharing platform 351 Execution Record Ledger 352 Ledger of Payment Terms 353 Responsibility demarcation ledger 354 Settlement Candidate Records Section 355 Payment Settlement Planning Record Department 356 Audit Records Ledger 357 Normalization rule storage unit 361 Performance Acquisition Department 362 Settlement Condition Acquisition Unit 363 Settlement candidate generation unit 364 Settlement and Evaluation Department 365 Settlement Planning Department 366 Settlement Execution Control Unit 367 Records Department

Claims

1. A payment settlement optimization system that determines the settlement of consideration between multiple corporations with respect to a case or processing unit executed collaboratively by AI agents or AI processing modules belonging to multiple corporations, A performance acquisition unit acquires execution performance data including case identifier, responsible company, processing results, resource usage record, result acceptance result, and deadline fulfillment result. A settlement conditions acquisition unit that acquires at least one of the consideration conditions, reduction conditions, retention conditions, incentive conditions, responsibility demarcation conditions, or settlement deadline conditions set for each of the aforementioned multiple corporations, A settlement candidate generation unit generates a plurality of settlement candidate items based on the execution performance data and the conditions obtained by the settlement condition acquisition unit, A settlement evaluation unit calculates an evaluation value for each of the aforementioned multiple settlement candidates based on at least one of the following: degree of conformity to contract terms, degree of consistency between income and expenditure, degree of reflection of contribution, degree of justification for reduction, degree of justification for reservation, or degree of suppression of dispute occurrence. A settlement plan determination unit that selects one of the multiple settlement candidates based on the evaluation value and confirms the settlement plan, A settlement execution control unit that generates at least one of the following based on the aforementioned settlement plan: billing data, allocation data, payment withholding data, or settlement confirmation notice data, A recording unit that stores the aforementioned execution performance data, the aforementioned multiple compensation settlement candidates, the aforementioned evaluation value, and the aforementioned compensation settlement plan as audit records, A payment settlement optimization system characterized by comprising the following features.

2. In the payment settlement optimization system described in claim 1, The aforementioned compensation conditions, reduction conditions, reservation conditions, incentive conditions, or liability demarcation conditions include at least one of the following: fixed compensation amount, usage-based charge rate, conditions fixed after acceptance of results, reduction rate in case of failure to meet SLA, burden ratio in case of re-execution, minimum guaranteed amount, success fee rate, or whether or not a reservation is required in case of dispute. A payment settlement optimization system characterized by the following features.

3. In the payment settlement optimization system according to claim 1 or 2, The settlement candidate generation unit generates at least one of the following: a fixed-amount allocation candidate, a contribution-proportional allocation candidate, an allocation candidate with quality result correction, a burden-sharing candidate according to the division of responsibility, or a candidate that reserves a portion of the consideration. A payment settlement optimization system characterized by the following features.

4. In the payment settlement optimization system according to any one of claims 1 to 3, The settlement evaluation unit reduces the evaluation value or excludes from the settlement candidates any of the multiple settlement candidates for which it detects a likelihood of breach of contract terms, budget overrun, inconsistency with the conditions for accepting results, inconsistency with the division of responsibilities, or likelihood of disputes arising. A payment settlement optimization system characterized by the following features.

5. In the payment settlement optimization system according to any one of claims 1 to 4, The settlement plan determination unit applies different settlement evaluation criteria depending on at least one of the following: the type of case, the confidentiality classification, the requesting corporation's priority policy, the recipient corporation's scope of responsibility, or the terms of continuing transactions. A payment settlement optimization system characterized by the following features.

6. In the payment settlement optimization system according to any one of claims 1 to 5, The settlement execution control unit, upon detecting at least one of the following: return of deliverables, occurrence of re-execution, deadline delay, resource overrun, approval expiration, or change in responsibility boundaries, issues a settlement recalculation request, making all or part of the settlement plan subject to recalculation. A payment settlement optimization system characterized by the following features.

7. In the payment settlement optimization system according to any one of claims 1 to 6, The recording unit generates a hash value corresponding to each payment settlement candidate or payment settlement plan, and stores it in association with the hash value corresponding to the preceding record. A payment settlement optimization system characterized by the following features.

8. In the payment settlement optimization system according to any one of claims 1 to 7, The settlement conditions acquisition unit or the settlement candidate generation unit utilizes at least one of the contract terms ledger, consideration ledger, settlement policy ledger, responsibility boundary ledger, or audit ledger, which are managed for each of multiple legal entities, after normalizing them based on version information, item correspondence information, or schema correspondence information. A payment settlement optimization system characterized by the following features.

Images