Information processing device, information processing method, and program

Abstract

To improve convenience for debt collectors. [Solution] An information processing device comprising: an acquisition unit that acquires an action history which is either or both of the debtor's repayment timing and their reaction to reminders; and a processing unit that sets a reminder pattern which is either or both of the reminder schedule and the reminder method for the debtor, based on the action history.

Description

Technical Field

[0001] The present invention relates to an information processing apparatus, an information processing method, and a program.

Background Art

[0002] Conventionally, a collection management system for managing debt collection has been disclosed (see, for example, Patent Document 1). The collection management system determines the collection means and the number of times for each debtor based on the cost for each collection means, debtor information regarding the debtor, and the budget available for implementing collection within a predetermined period.

Prior Art Documents

Patent Documents

[0003]

Patent Document 1

Summary of the Invention

Problems to be Solved by the Invention

[0004] However, the management of debt collection may not always be sufficiently convenient for the collector (operator).

[0005] The present invention has been made in consideration of such circumstances, and one of its objects is to provide an information processing apparatus, an information processing method, and a program that can improve the convenience for the collector.

Means for Solving the Problems

[0006] One aspect of the present invention is an information processing apparatus including an acquisition unit that acquires an action history that is one or both of the repayment timing of a debtor and the reaction to collection, and a processing unit that sets a collection mode that is one or both of the collection schedule and the collection method for the debtor based on the action history.

Effects of the Invention

[0007] According to one aspect of the present invention, it is possible to provide an information processing device, an information processing method, and a program that can improve the convenience of debt collectors. [Brief explanation of the drawing]

[0008] [Figure 1] This figure shows an example of the configuration of a support system for debt collection. [Figure 2] This figure shows an example of the functional configuration of the support device 10. [Figure 3] This figure shows an example of the contents of the debt information 52 stored in the memory unit 50. [Figure 4] This figure shows an example of the contents of behavioral history information 53. [Figure 5] This figure shows an example of the contents of the base flow. [Figure 6] This is a diagram to explain the contents of the adjustment flow. [Figure 7] This diagram illustrates the adjustment flow for the most effective repayment cycle. [Figure 8] This diagram illustrates the flow chart for adjusting the effectiveness of debt collection methods. [Figure 9] This flowchart shows an example of the processing flow performed by the support device 10. [Figure 10] This flowchart shows another example of the processing flow performed by the support device 10. [Figure 11] This figure shows another example of the configuration of a support system for debt collection. [Modes for carrying out the invention]

[0009] Hereinafter, embodiments of the information processing apparatus, information processing method, and program of the present invention will be described with reference to the drawings.

[0010] [Support System] Figure 1 shows an example of the configuration of a support system for debt collection. Support system 1 includes, for example, a support device (information processing device) 10, an operator terminal device (debtor's terminal device) 200, and a user terminal device (debtor's terminal device) 300. The support device 10, the operator terminal device 200, and the user terminal device 300 can communicate with each other via a network NW. The network NW includes, for example, the Internet, LAN (Local Area Network), wireless base stations, provider equipment, etc.

[0011] [Operator terminal device] The operator terminal device 200 is a terminal device used by the operator. The operator uses the operator terminal device 200 to identify the debtors to be pursued and to set up collection efforts to be made to those debtors using predetermined collection methods. The operator terminal device 200 displays various information, which will be described later, on its display unit.

[0012] [User terminal device] The user terminal device 300 is, for example, the terminal device of a debtor who has a debt. The user terminal device 300 is, for example, a smartphone or a personal computer. The user terminal device 300 can view various information managed by the support device 10 and register various information in cooperation with the support device 10.

[0013] [Support device] FIG. 2 is a diagram showing an example of the functional configuration of the support device 10. The support device 10 includes, for example, a communication unit 12, a display control unit 14, an information management unit 16, an information processing unit 18, and a storage unit 50. Components other than the communication unit 12 and the storage unit 50 are realized, for example, when a hardware processor such as a CPU executes a program (software). Some or all of these components may be realized by hardware (including a circuit unit; circuitry) such as an LSI (Large Scale Integration), an ASIC (Application Specific Integrated Circuit), an FPGA (Field-Programmable Gate Array), or a GPU (Graphics Processing Unit), or may be realized by cooperation between software and hardware. The program may be stored in advance in a storage device (a storage device having a non-transitory storage medium) such as an HDD (Hard Disk Drive) or a flash memory, or may be stored in a removable storage medium (a non-transitory storage medium) such as a DVD or a CD-ROM, and may be installed in the storage device when the storage medium is mounted on a drive device. The information management unit 16 is an example of an "acquisition unit". The information processing unit 18 is an example of a "processing unit".

[0014] The storage unit 50 is an HDD, a flash memory, a RAM (Random Access Memory), or the like. The storage unit 50 may be a NAS (Network Attached Storage) device accessible by the support device 10 via a network. The storage unit 50 stores, for example, creditor information 52, action history information 53, operator information 54 (for example, name and attendance time) which is information about an operator, reminder flow information 56, personalization information 58, content information 70, and the like. The content information 70 is, for example, information for displaying content related to reminders or an interface screen. Details of other information will be described later.

[0015] The communication unit 110 is a communication interface for connecting to the network NW. The communication unit 110 is, for example, a network interface card.

[0016] The display control unit 14 provides, for example, various interface screens of the content information 70 to the operator terminal device 200, the user terminal device 300, etc. The display control unit 14 appropriately reads out necessary content from the content information 70 and provides it to the target device or the like.

[0017] The information management unit 16 acquires various information from the operator terminal device 200, the user terminal device 300, etc., and manages the acquired information. The information management unit 16 manages the action history information 53.

[0018] The information processing unit 18 performs various information processes using the information stored in the storage unit 50 and the information acquired from the operator terminal device 200, the user terminal device 300, etc. The information processing unit 18 generates, for example, the reminder flow information 56. Details of the processing of the information processing unit 18 will be described later.

[0019] [Claim Information] FIG. 3 is a diagram showing an example of the content of the claim information 52 stored in the storage unit 50. The claim information 52 is information in which a customer number (debtor number), a corporate name, a name, the total amount of claims, the scheduled reminder date and time, a person in charge (operator), a reminder flow (described later), information on the applicability of the personalized reminder flow (described later), etc. are associated. Different information from the above information may be associated with the claim information 52.

[0020] The debt collection flow is managed as debt collection flow information 56. The debt collection flow is a combination of debt collection methods and timing set by the operator. For example, it is information that specifies a debt collection schedule, such as making a debt collection at a specific time using a designated debt collection method on the Xth day of the overdue period. The operator can set the debt collection flow via the operator terminal device 200 on the interface screen provided by the support device 10. Debt collection methods include email, SMS, automated phone calls, and manual phone calls. Email, SMS, automated phone calls, etc., are carried out automatically by the support device 10 based on the schedule set in the debt collection flow.

[0021] The suitability of the personalized debt collection flow is information indicating the suitability of the mode in which the debt collection method and timing are set based on the debtor's behavioral history information 53, as will be described later.

[0022] [Activity History Information] Figure 4 shows an example of the contents of the behavioral history information 53. The behavioral history information 53 is information that associates the date and time a demand for payment was made, the method of demand for payment, and the debtor's reaction to the demand for payment. For example, for email, whether the debtor clicked on the URL in the body of the email and viewed the information provided by the URL, or whether they did not click on it, is stored as behavioral history information 53. For example, for SMS, whether the debtor clicked on the URL in the body of the SMS and viewed the information provided by the URL, or whether they did not click on it, is stored as behavioral history information 53. For example, whether the debtor spoke with an operator via automated or manual telephone, or whether the call was unsuccessful, is stored as behavioral history information 53.

[0023] [Personalized Information] Personalized information 58 is information that indicates the setting criteria for the collection method and timing of collection when a personalized collection flow is applied. The personalized collection flow is specified by the operator. For example, the operator can set whether to apply a personalized collection flow or a collection flow customized by the operator on the interface screen provided by the support device 10 to the operator terminal device 200. Personalized information 58 includes, for example, a base flow, an oversight adjustment flow, an adjustment flow for the most effective repayment cycle, and an adjustment flow for the effectiveness of the collection method (details are described later).

[0024] [Use of behavioral history information] The support device 10 sets a debt collection schedule and a debt collection method, which is either or both, based on the debtor's behavioral history, which is either the timing of repayments or their reaction to debt collection notices. For example, the following are set: [Accidental Adjustment Flow], [Most Effective Repayment Cycle Adjustment Flow], and [Effectiveness of Debt Collection Method Adjustment Flow], which will be described later.

[0025] The support device 10, for example, may skip a reminder by adjusting the accidental adjustment flow, change the timing of reminders by adjusting the most effective repayment cycle flow, or set the method of reminder by adjusting the effectiveness of the reminder means flow.

[0026] The support device 10 issues a reminder to the debtor based on a pre-set base flow, and then determines the type of reminder based on the action history information 53 for the base reminder flow. If action history information 53 for the debtor's claim (or other claims of the debtor) exists, that action history information 53 may be used, or action history information 53 for another reminder flow (for example, a reminder flow set by the operator) may be used instead of the base flow.

[0027] [Base flow] Figure 5 shows an example of the contents of the base flow. The base flow is a flow in which an automated phone call is made once a day starting one day after the delinquency, an email and SMS are sent once every three days starting four days later, and a postal mail is sent every 25 days starting 25 days later. For example, for debtors for whom there is no behavioral history information 53 up to this point, the base flow is set when a personalized collection flow is set. After the base flow is implemented for a predetermined period (e.g., 30 or 60 days), the collection flow is adjusted by the oversight adjustment flow, the most effective repayment cycle adjustment flow, and the effectiveness of collection methods adjustment flow. The base flow and the flows described later are collection methods that comply with laws, industry customs, and other rules, and take into consideration the debtor's life.

[0028] [Accidental Adjustment Flow] The support device 10 skips the execution of pre-set collection notices for a predetermined period if the debtors have previously paid off debts that meet predetermined criteria. For example, the support device 10 determines that the predetermined criteria are met if the proportion of debts that have been paid off and whose final repayment date is earlier than the date on which the debt was entrusted to or transferred to another party (with a predetermined number of days added to that date) is equal to or greater than a threshold.

[0029] Figure 6 is a diagram illustrating the contents of the accidental adjustment flow. In the accidental adjustment flow, if the conditions are met, the reminder schedule for 7 days from the date set for a reminder will be skipped. For example, if the reminder date is October 1st, reminders scheduled from October 1st to October 7th will be skipped, and reminders scheduled from October 8th onwards will be implemented.

[0030] The condition is that the calculation result in (1) below, applied to fully repaid claims, is equal to or greater than the threshold (for example, 0.8 or higher). The first type of claim is the number of fully repaid claims that satisfy the condition (final repayment date ≤ date of acceptance or transfer date plus a specified date (for example, 7)). For example, if the final repayment date is October 10th and the date the claim was accepted or transferred to another party is October 5th, adding 7 to October 5th results in October 12th, so it is a first type of claim. For example, if the final repayment date is October 10th and the date the claim was accepted or transferred to another party is October 2nd, adding 7 to October 2nd results in October 9th, so it is not a first type of claim. (1) Calculation result = Number of first-class claims / Total number of fully repaid claims

[0031] [Adjustment flow for the most effective repayment cycle] Based on the behavioral history information 53, the support device 10 sets the timing for issuing a reminder to the debtor at a time within the set period when the debtor is most likely to make a repayment. The support device 10 aggregates the repayment dates in the behavioral history for each divided period, which is a monthly period divided according to a predetermined standard, and issues the reminder in the divided period when the number of repayment dates in that period meets the standard.

[0032] Figure 7 is a diagram illustrating the adjustment flow for the most effective repayment cycle. The adjustment flow for the most effective repayment cycle is set up so that reminders are prioritized or actively sent during periods when the conditions are met. For example, the number of automated phone calls per day for repayment cycles that meet the conditions is changed to three times a day. For example, the first automated call is made between 10:30 and 12:30, and then automated calls are made at two-hour intervals thereafter.

[0033] The period that satisfies the conditions is one in which the repayment dates on the calendar are aggregated for all repayment data in the behavioral history information 53, and the total number of repayments exceeds a threshold (e.g., 40% or more). The period is a divided period in which one month is divided based on days ending in 5 or 10, for example, from the 1st to the 5th, the 6th to the 10th, the 16th to the 15th, the 16th to the 20th, the 21st to the 25th, the 26th to the 31st, etc. For example, if the number of repayments is less than two, the evaluation of the most effective repayment period does not need to be performed.

[0034] [Adjustment flow for the effectiveness of debt collection methods] Based on the behavioral history information 53, the support device 10 sets the channel with the highest probability of being effective for the debtor among multiple collection channels as the collection channel. The support device 10 prioritizes either the first channel or the second channel for collection based on an index obtained by statistically processing the response rate to the first channel and the response rate to the second channel. The first channel is, for example, SMS, and the second channel is, for example, email.

[0035] Figure 8 is a diagram illustrating the content of the adjustment flow for the effectiveness of debt collection methods. The adjustment flow for the effectiveness of debt collection methods is an adjustment flow in which the debt collection method is changed according to the effectiveness indicator. For example, if there has been at least one instance of sending an email and an SMS during a predetermined period (e.g., 6 months) prior to the debt collection setting date, the effectiveness of each method is evaluated.

[0036] The effectiveness of emails and SMS messages is calculated using the following formula: Email effectiveness = (Number of times the login URL in the email body was clicked) / (Number of times the email was sent) SMS effectiveness = (Number of times the login URL in the SMS message was clicked) / (Number of times SMS messages were sent) The difference in effectiveness between the two channels is calculated using the following formula. Difference in effectiveness (metric) = (Email effectiveness) - (SMS effectiveness) Please note that even if you click the login URL multiple times per reminder, it will only be counted as one click.

[0037] Depending on the difference in effectiveness, the collection methods will be changed as follows: 0.5 < Difference: SMS messages for the past 30 days will be overwritten by emails. 0 ≤ difference ≤ 0.5: SMS messages from the past 10 days will be overwritten by emails. -0.5 ≤ difference < 0: Emails for the past 10 days will be overwritten by SMS. If the difference is ≤ -0.5, emails will be overwritten by SMS messages for the past 30 days.

[0038] For example, the support device 10 determines, in this order, whether the action history information 53 regarding the base flow for debt collection corresponds to the [accidental adjustment flow], the [most effective repayment cycle adjustment flow], or the [adjustment flow for the effectiveness of the debt collection method]. If the support device 10 corresponds to the [accidental adjustment flow], it skips the debt collection. If the support device 10 corresponds to the [most effective repayment cycle adjustment flow] or the [adjustment flow for the effectiveness of the debt collection method], it sets the corresponding adjustment flow in addition to the [accidental adjustment flow]. The support device 10 may apply the [accidental adjustment flow], the [most effective repayment cycle adjustment flow], or the [adjustment flow for the effectiveness of the debt collection method] to the base flow, or it may apply the [accidental adjustment flow], the [most effective repayment cycle adjustment flow], or the [adjustment flow for the effectiveness of the debt collection method] to a debt collection flow different from the base flow. A different debt collection flow may be a debt collection flow that is set in advance by the system and is different from the base flow, or it may be a debt collection flow set by the operator.

[0039] For example, if the [Accidental Adjustment Flow] applies, the [Accidental Adjustment Flow] will take precedence and be applied. If repayment is not made during the period of the [Accidental Adjustment Flow], a reassessment will be conducted. Alternatively, if the assessment prior to the implementation of the [Accidental Adjustment Flow] found that either or both of the [Most Effective Repayment Cycle Adjustment Flow] and / or the [Adjustment Flow for the Effectiveness of Collection Methods] apply, the applicable flow will be applied and collection will be initiated.

[0040] Furthermore, if the debt collection period is extended, the prescribed adjustment flow does not need to be applied. The prescribed adjustment flows include, for example, the [accidental adjustment flow] and the [adjustment flow for the effectiveness of debt collection methods]. An extension of debt collection occurs, for example, when a new debt collection flow is set up at predetermined intervals, or when debt collection is resumed because promises such as payment or contact are not fulfilled. In these cases, the adjustment flow is not applied, and the base flow is automatically set. For example, if the debt collection flow is manually set to a personalized debt collection flow, or if a new debt is registered after full repayment or after the debt has been removed from management, this does not constitute an extension of debt collection, and therefore the prescribed adjustment flow will be applied.

[0041] [Flowchart (1)] Figure 9 is a flowchart showing an example of the processing flow executed by the support device 10. First, the support device 10 determines whether a personalized reminder flow has been set (S100). If a personalized reminder flow has been set, the support device 10 determines whether past behavior history information 53 exists (S102). If past behavior history information 53 exists, the process in S104 is skipped. If past behavior history information 53 does not exist, the support device 10 sets the base flow (S104). This completes the processing of one routine in this flowchart.

[0042] [Flowchart (2)] Figure 10 is a flowchart showing another example of the processing flow performed by the support device 10. First, the support device 10 determines whether or not it is a judgment timing (S200). For example, if past behavior history information 53 exists, the timing at which a personalized reminder flow is set is the judgment timing. For example, if past behavior history information 53 does not exist, the timing at which a reminder flow based on a predetermined period is implemented is the judgment timing.

[0043] If it is the timing for a determination, the support device 10 refers to the behavior history information 53 (S202) and determines whether the behavior history information 53 meets the conditions included in the personalized reminder flow (S204). If the conditions included in the personalized reminder flow are not met, for example, the base flow may continue, or reminder may be executed according to the reminder flow that the operator has previously set to be executed after the base flow.

[0044] If the conditions included in the personalized reminder flow are met, the support device 10 applies the personalized reminder flow that meets the conditions and performs reminder (S206). For example, as described above, the personalized reminder flow that meets the conditions is applied to the base flow, or the personalized reminder flow that meets the conditions is applied to a pre-configured reminder flow. This completes the processing of one routine in this flowchart.

[0045] As described above, the support device 10 can improve the operator's convenience by setting a debt collection method, which may be either a schedule for debt collection or the method of debt collection, or both, based on the user's activity history. For example, the support device 10 can improve the collection rate by carrying out debt collection in the set debt collection method.

[0046] Specifically, the support device 10 can efficiently elicit repayment behavior and improve the collection rate by conducting debt collection at the optimal timing and through channels tailored to the debtor's behavioral characteristics. Furthermore, by automating responses that are tailored to the customer's situation while complying with laws and regulations, the support device 10 can maintain long-term customer relationships without damaging customer trust even at negative points such as delinquency. In addition, the support device 10 can set up a flow that improves the collection rate without relying on the individual for debt collection work. Because the support device 10 performs "customer identification" and "optimization of approach," which previously relied on the experience and intuition of operators, based on behavioral history, the quality of debt collection work across the entire organization is standardized and improved. Moreover, the above automation can automatically reduce ineffective and wasteful debt collection actions and concentrate resources on the channels that are most likely to produce a response, thereby reducing collection costs.

[0047] [Real-time optimization] The support device 10 may generate a flow based on the behavior history information 53 using a learning model. The learning model is a model that has learned from learning data. The learning data includes, for example, past "states," past "actions," and "rewards" for the above-mentioned "states" and "actions." The learning model learns the correlation between "states" and "actions" and the corresponding "rewards," and is trained to output the "action" that maximizes the "reward" depending on the "state." For example, based on the above-learned correlation, the learning model predicts in real time the optimal debt collection method "action" that maximizes the "reward" from a given "state." For example, "state" is the customer's status, such as the number of days overdue or their response to past debt collection attempts. "Action" is a debt collection action, such as the debt collection channel, timing of the debt collection, and the wording of the debt collection attempt. "Reward" is the collection rate or the debtor's positive reaction.

[0048] As described above, the support device 10 can make more optimal debt collection efforts for debtors by using the debt collection methods predicted in real time by the learning model.

[0049] [Sentiment analysis] The support device 10 may predict the debtor's emotions by analyzing chats with the debtor, messages from the debtor, call records (text obtained by speech recognition), etc., using natural language processing technology. This analysis is performed using known methods. If the support device 10 determines that the debtor has negative emotions such as anger or confusion, it may automatically temporarily lower the debt collection level or escalate the case to a responsible operator (an experienced operator). Lowering the debt collection level means easing the number and frequency of debt collection attempts, or using a less burdensome method of debt collection, such as email or SMS instead of phone calls. Escalation is carried out, for example, by providing the debtor's information to the operator terminal device 200 of the responsible operator.

[0050] As described above, the support device 10 can predict the debtor's emotions and take appropriate action according to the predicted emotions, thereby providing empathetic support to the debtor.

[0051] [Use of Generative AI] Figure 11 shows another example of the configuration of a support system for debt collection. Support system 1A includes, for example, a generating AI 400 in addition to the functional configuration of support system 1. Support device 10 and generating AI 400 can communicate via a network NW. Generating AI 400 is a large-scale language model (LLM) that has been trained to learn from a large amount of information, such as information published or provided on the network NW, and to output information corresponding to the input information and requests.

[0052] [Personalized communication] The support device 10 may use the generating AI 400 to dynamically generate debtor personas and personalize communication. The generating AI 400 uses behavioral history information 53 to create detailed "personas" in text, inferring the debtor's personality and situation, not just simple categories (e.g., "accidental adjustment flow"). Based on these personas, the generating AI 400 completely individually generates the wording and tone of reminder messages for each person. For example, in addition to personalization such as "which channel to send it through and when," the generating AI 400 also generates the "quality" of communication, such as "what to say and how to say it." For example, the generating AI 400 generates a persona such as "a busy business owner who prefers polite but concise language," and creates email text tailored to that persona.

[0053] For example, the Generating AI 400 learns combinations of behavioral history information 53 (and debtor attributes) and personas as training data, and learns to output a persona corresponding to the behavioral history information 53 when behavioral history information 53 (and debtor attributes) is input. Furthermore, the Generating AI 400 learns personas and corresponding text, and learns to output text corresponding to the persona and delinquency status when a persona and delinquency status are input.

[0054] The Generating AI400 generates a detailed persona in natural language from multifaceted data such as past repayment patterns, response speed to reminders, and types of services used, describing the customer as "a cautious individual who tends to pay in a lump sum after payday" or "a planned type who prefers frequent reminders." Based on the generated persona and the current delinquency status, the Generating AI400 generates email and SMS messages from scratch. The Generating AI400 automatically generates the most effective expressions that resonate with the recipient (e.g., empathetic tone, approach that highlights benefits), which is impossible with templates.

[0055] The Generating AI 400 may generate messages based on the debtor's nationality and place of residence, taking into account the most appropriate polite expressions and phrases for that cultural sphere, as well as the official language of the country of nationality. The Generating AI 400 learns information such as the debtor's nationality and place of residence, along with messages corresponding to that information, as training data, and is trained to output a message corresponding to that information when the above information is input.

[0056] Through the above process, debt collection communications evolve from "automated administrative communications" to "personalized, human-centered dialogues." This reduces the debtor's psychological resistance to payment and encourages voluntary repayment behavior without damaging the customer relationship. The support device 10 can elevate each of the adjustment flows described above to a higher level of action.

[0057] [Automated negotiation and consensus building] The generating AI 400 may function as an agent for automated negotiation and agreement on payment plans. When a debtor receives a reminder message (e.g., SMS) and transitions to a chat screen via the URL provided in the SMS, the AI ​​agent automatically responds to inquiries and negotiations regarding payments in a conversational format, completing the process until a realistic payment plan is agreed upon. The AI ​​agent communicates with the generating AI 400 and provides services to the debtor via the support device 10.

[0058] The Generating AI 400 understands requests from debtors in natural language, such as "I can't pay in a lump sum" or "I need to wait until payday," and proposes and negotiates the most suitable installment payment or payment deadline extension plan within the company's rules. For example, the Generating AI 400 learns information that associates company rules, debtors' requests, debtors' behavioral history information 53, past conversations, and proposals or negotiations as training data. When a debtor makes a request, it outputs a proposal or negotiation based on the request, the debtors' behavioral history information 53, and past conversations.

[0059] The generating AI 400 understands, for example, vague requests or inquiries from debtors, and, while comparing them with the company's collection policy (e.g., installments up to 3 times, deferrals up to 30 days), presents multiple payment options and outputs text to facilitate agreement through dialogue. If a payment plan is agreed upon through dialogue with the debtor, the generating AI 400 feeds back the details (e.g., half payment next Friday) to the support device 10, which then automatically pauses or modifies the ongoing debt collection flow.

[0060] If the generating AI 400 detects strong dissatisfaction or anger from the debtor based on the content of the conversation, it may automatically interrupt the conversation and immediately hand it over to a human operator (escalate the situation) to manage the risk. For example, the generating AI 400 may escalate the situation to the support device 10, and the support device 10 may escalate it to the operator terminal device 200.

[0061] As described above, debtors can discuss payment options 24 / 7, 365 days a year, at their own convenience, without the need for an operator. This frees operators from simple negotiation tasks, allowing them to focus on more complex cases, dramatically improving operational efficiency. For example, the above process optimizes the creation of contact points with debtors, and the AI400 generates those contact points to translate into concrete collection results, thereby improving the collection rate.

[0062] [Health support] The Generating AI 400 may proactively detect and support the recovery of "potential defaulters." The Generating AI 400 may use machine learning to analyze a vast amount of past behavioral history information 53, pre-default debtor behavioral data (e.g., repayment status, service usage frequency, payment cycle, etc.), and debtor attributes to predict with high accuracy debtors who show signs of default "before" default occurs. The Generating AI 400 then automatically proposes positive engagement (e.g., suggestions for advantageous payment methods, support guidance) rather than debt collection. For example, the Generating AI 400 takes on the role of preventative action "before default occurs." It captures "signs" such as a decrease in the debtor's service usage frequency or slight disruptions in the payment cycle, and takes proactive measures before the problem becomes serious.

[0063] The Generating AI400 has a delinquency risk scoring function. The Generating AI400 monitors the behavioral data of all debtors in real time and constantly calculates a score indicating the likelihood of delinquency. The Generating AI400 automatically lists customers whose scores have suddenly risen as "high-risk." For the listed customers, the Generating AI400 automatically proposes or selects and executes positive actions that are completely different from debt collection (for example, "notification of changing to revolving credit," "notification of the benefits of consolidation payment," "guidance to a financial planning consultation service," etc.) according to their risk level and persona. For example, the Generating AI400 is trained to output a score corresponding to behavioral data when behavioral data is input, and further trained to output a positive action from the services it offers when either or both of the risk level (e.g., score) and / or persona of a high-risk debtor are input. As a result, the Generating AI400 outputs scores and positive actions according to the input information.

[0064] The generating AI400 may continuously learn how effective the proactive actions (positive actions) it has taken have been in preventing subsequent delinquency, and may self-evolve the accuracy of its interventions. For example, if the generating AI400 proposes a first positive action to a debtor and delinquency is suppressed, it learns that the proposal of the first positive action is effective and will continue to propose that first positive action thereafter. For example, if the generating AI400 proposes a second positive action to a debtor and delinquency is not suppressed, it learns that the proposal of the second positive action is ineffective and will not propose that second positive action thereafter.

[0065] The above process can fundamentally reduce the number of debt collection tasks that occur. Furthermore, by offering support from the creditor's side before debtors fall into financial difficulties, customer loyalty can be improved and the company's brand image can be enhanced.

[0066] According to the embodiments described above, the support device 10 can improve the operator's convenience by setting a debt collection schedule and a debt collection method, which are either or both, based on the debtor's behavioral history, which are either or both, the debtor's repayment timing and their reaction to debt collection.

[0067] Although embodiments for carrying out the present invention have been described above using examples, the present invention is not limited in any way to these embodiments, and various modifications and substitutions can be made without departing from the spirit of the present invention. [Explanation of Symbols]

[0068] 1. 1A Support System 10 Support equipment 14 Display Control Unit 16 Information Management Department 18 Information Processing Department 50 Storage section 52. Debt Information 53. Activity History Information 58 Personalized Information 70 Content Information 200 Operator Terminal Device 300 User terminal devices 400 Generation AI

Claims

1. An acquisition unit that acquires behavioral history, which is either or both of the debtor's repayment timing and their reaction to reminders, The system includes a processing unit that sets a debt collection method, which is either or both, a schedule for debt collection and a method for debt collection, based on the aforementioned behavioral history. The aforementioned processing unit, If the debtor has previously paid off a debt that meets the specified criteria, the predetermined collection period will be skipped. Of the aforementioned fully repaid claims, the predetermined criteria are met if the proportion of claims whose final repayment date is earlier than the date on which the claim was entrusted to or transferred to another party (with a predetermined number of days added to that date) is equal to or greater than the threshold. Information processing device.

2. The processing unit sets one or more of the following reminder modes based on the behavior history: skipping the reminder, changing the timing of the reminder, and the method of the reminder. The information processing apparatus according to claim 1.

3. The processing unit, after making the debt collection request to the debtor based on a pre-configured base debt collection flow, determines the debt collection method based on the action history in relation to the base debt collection flow. The information processing apparatus according to claim 1.

4. The aforementioned processing unit, Based on the aforementioned behavioral history, the timing of the debtor's repayment is set to the timing of the debt collection notice, where the probability of the debtor making a repayment within the specified period is high. The information processing apparatus according to claim 1 or 2.

5. The aforementioned processing unit, For each period of the month divided according to a predetermined standard, the repayment dates in the aforementioned activity history are totaled, and the aforementioned reminder is sent for the period in which the number of repayment dates in that period meets the standard. The information processing apparatus according to claim 4.

6. The aforementioned processing unit, Based on the aforementioned behavioral history, the channel with the highest probability of being effective for the debtor among multiple collection channels is set as the collection channel. The information processing apparatus according to claim 1 or 2.

7. The aforementioned processing unit, Based on an index obtained by statistically processing the response rate to the first channel of reminders and the response rate to the second channel of reminders, the reminders are given priority to the first channel or the preceding second channel. The information processing apparatus according to claim 6.

8. The first channel is SMS, The aforementioned second channel is email. The information processing apparatus according to claim 7.

9. Computers By obtaining a history of the debtor's repayment timing and / or their reaction to reminders, Based on the aforementioned behavioral history, a debt collection method is set, which is either a schedule for debt collection and / or the method of debt collection. If the debtor has previously paid off a debt that meets the specified criteria, the predetermined collection period will be skipped. Of the aforementioned fully repaid claims, the predetermined criteria are met if the proportion of claims whose final repayment date is earlier than the date on which the claim was entrusted to or transferred to another party (with a predetermined number of days added to that date) is equal to or greater than the threshold. Information processing methods.

10. On the computer, By obtaining a history of the debtor's repayment timing and / or their reaction to reminders, Based on the aforementioned behavioral history, the debt collection method is set, which is either a schedule for debt collection and / or the method of debt collection. If the debtor has previously paid off a debt that meets the specified criteria, the predetermined collection period will be skipped. Of the aforementioned fully repaid claims, if the proportion of claims whose final repayment date is earlier than the date on which the claim was entrusted to or transferred to another party (with a specified number of days added to that date) is equal to or greater than a threshold, the predetermined criteria will be deemed to be met. A program for that purpose.

Images