A user portrait-based accurate fee urging intelligent analysis method and system thereof

Abstract

The application discloses a kind of based on user portrait's accurate fee urging intelligent analysis method and system, specifically related to intelligent fee urging management technical field, the method obtains basic data and establishes fee urging analysis transaction, determines effective arrear bill and arrear feature information;According to historical payment behavior and fee urging touch behavior, generate user original payment portrait, and determine user natural payment baseline by natural payment sample;Combining fee urging priority comprehensive quantity and touch adaptation comprehensive quantity generates differentiated fee urging strategy and strategy intervention record;After fee urging execution, according to the relationship of repayment time point, touch intensity before repayment and front response state determine fee urging intervention influence quantity, and accordingly execute portrait update, strategy sample save or delay update processing, form fee urging closed loop state.

Description

Technical Field

[0001] This invention relates to the field of intelligent payment collection management technology, and more specifically, to a precise intelligent analysis method and system for payment collection based on user profiles. Background Technology

[0002] As property management fee collection gradually moves online, property management companies typically manage bills, payment records, reminder contact records, and payment results through property fee collection systems, user service platforms, SMS platforms, application messaging platforms, outbound calling systems, and online payment platforms. Existing reminder systems generally filter recipients based on outstanding amounts, overdue days, historical payment history, contact information validity, and past contact feedback, and then arrange reminders via SMS, application messaging, outbound calls, or human customer service. Some systems also incorporate user profiling or tagging rules to determine reminder priorities or contact methods based on users' historical payment habits, payment channel preferences, and reminder response rates, thereby improving the efficiency of reminder task allocation.

[0003] However, in actual property management fee collection scenarios, the reasons why users complete payments do not necessarily directly reflect their original payment habits. Some payments may occur around the user's natural payment time, while others are made after multiple reminders, outbound calls, intervention from human customer service, or multiple channels of contact. Still others may be affected by factors such as fee disputes, unclear bill matching, or abnormal contact information. Existing solutions typically focus more on whether a payment has been made and after which contact method it occurred, easily incorporating payments heavily influenced by aggressive collection actions directly into the user profile. This can lead to deviations in subsequent collection priority, contact intensity, and contact method judgments. Therefore, it remains necessary to provide a precise intelligent analysis method and system for property management fee collection scenarios to improve the accuracy of updating user payment profiles and generating differentiated collection strategies. Summary of the Invention

[0004] In order to overcome the above-mentioned defects of the prior art, embodiments of the present invention provide a precise intelligent analysis method and system for debt collection based on user profiles to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution: A precise payment collection intelligent analysis method based on user profiles includes the following steps: Acquire basic data, establish a collection analysis process, and identify valid overdue bills and their corresponding overdue characteristic information; For users with valid overdue bills, an original payment profile is generated based on historical payment behavior and reminder contact behavior. Natural payment samples are then selected from historical payment records, and the user's natural payment baseline is determined based on the natural payment samples. Based on the valid overdue bills, overdue feature information, user's original payment profile, and user's natural payment baseline, the collection priority comprehensive quantity is determined, and the corresponding reach adaptation comprehensive quantity is determined for different collection reach methods. Based on the collection priority comprehensive quantity and the reach adaptation comprehensive quantity corresponding to each collection reach method, a differentiated collection strategy is generated, and a strategy intervention record is formed at the same time. After the collection task is executed, the collection reach feedback information and payment feedback information are obtained. Based on the strategy intervention record, the relationship between payment time, the intensity of reach before payment and the pre-repayment response status, the impact of the collection intervention is determined. The credibility status of the feedback is determined based on the impact of the payment reminder intervention, and the user payment profile is updated, the effect sample of the payment reminder strategy is saved or the update is temporarily suspended according to the credibility status of the feedback, thus forming a closed loop of payment reminder.

[0006] In a preferred embodiment, the basic data includes pending bills, historical payment records, and historical collection records, and includes at least one of contact feedback records, payment channel usage records, and user service records; when establishing a collection analysis transaction, user identifier, house identifier, and payment cycle are used as the main association fields to associate the bill data, payment data, collection contact data, and service interaction data of the same user in different data sources, and generate a collection transaction number.

[0007] In a preferred embodiment, when determining a valid overdue bill, bill records for the same user, the same house, and the same billing cycle are processed for consistency. Bills with an amount due greater than the amount paid and a current date later than the due date are determined to be valid overdue bills. Bills that do not meet the above conditions are identified as invalid overdue bills and are not included in the current collection process. The overdue characteristic information includes at least one of the following: overdue amount, number of overdue days, number of consecutive overdue cycles, cumulative number of overdue payments, and the interval between the most recent payment.

[0008] In a preferred embodiment, the original user payment profile is generated based on the user's payment behavior and reminder contact behavior within a preset historical time window. The original user payment profile includes at least one of the following: number of on-time payments, number of overdue payments, distribution of natural payment time, number of payments after reminders, historical SMS response status, historical phone call connection status, historical payment channel preference, and historical payment interval.

[0009] In a preferred embodiment, when screening natural payment samples, payment records formed within the billing cycle that did not involve outbound phone calls, customer service follow-ups, continuous outbound calls, or strong scripted reminders are used as natural payment samples. Low-intervention payment records formed only after receiving due date reminders, SMS reminders, or application message reminders are also used as natural payment samples. Payment records formed after strong collection efforts have occurred are excluded from the user's natural payment baseline sample.

[0010] In a preferred embodiment, when determining the user's natural payment baseline, the number of days the payment time is offset from the due date is used as the analysis object. When the number of natural payment samples reaches a preset sample size threshold, the user's natural payment baseline is determined based on the median value of the offset days in the natural payment samples and the allowable range of natural payments. The allowable range of natural payments is determined based on the dispersion of the offset days, or by the interval between the upper and lower quantile values ​​of the offset days. When the number of natural payment samples is less than the preset sample size threshold, the user's natural payment baseline is marked as a low-confidence baseline.

[0011] In a preferred embodiment, when determining the overall priority of collection, it is determined based on at least a portion of the following: level of overdue amount, level of overdue days, level of consecutive overdue period, level of historical collection response, and level of impact of service dispute. When there are unclosed fee disputes, unclosed complaints, or unclosed matters under manual verification, the level of impact of service dispute is set to the highest preset level, and the corresponding collection analysis matter is transferred to the dispute verification and processing process, while the regular collection intensity escalation operation is suspended.

[0012] In a preferred embodiment, when determining the overall reach matching quantity, corresponding overall reach matching quantities are generated for different debt collection reach methods. Based on the historical reach success rate, post-reach payment interval, payment gateway response status, user's commonly used payment channels, contact information validity status, reach frequency limit status, and user's natural payment baseline, the corresponding debt collection reach methods are classified into high-fit reach methods, medium-fit reach methods, low-fit reach methods, or unavailable reach methods. Among them, unavailable reach methods are not included in the current debt collection reach order, and low-fit reach methods are only used as alternative methods when other reach methods are unavailable or require manual review.

[0013] In a preferred embodiment, the strategy intervention record includes at least the collection transaction number, contact method, contact order, contact time, number of contacts, intensity of the communication, whether it includes human intervention, whether it includes a payment link, whether it is within the user's historical active payment period, and contact level. When determining the impact of the collection intervention, the collection time is determined based on the relationship between the current collection time and the user's natural payment baseline; the pre-collection contact intensity is determined based on the actual collection contact intensity experienced by the user before completing the payment; and the pre-response status is determined based on whether there was an effective response behavior before the strong contact occurred. This includes at least one of the following: outbound phone calls, customer service follow-up, continuous outbound calls, or offline notifications. If the current payment is close to the user's natural payment baseline and the outreach intensity before the payment is low, it is determined to be a low-intervention-impact state. If the current payment deviates from the user's natural payment baseline, the outreach intensity before the payment is high, and there was no effective response behavior before the strong outreach occurred, it is determined to be a high-intervention-impact state, and the corresponding payment result is saved as a sample of the effectiveness of the collection strategy, and is not directly used to update the user's payment profile. If the contribution of multiple channels in the same collection transaction is unclear, it is determined to be in an attribution pending verification state and the update is temporarily suspended.

[0014] In a preferred embodiment, the system includes: The data processing unit is used to acquire basic data, establish collection analysis transactions, identify valid overdue bills, and generate overdue characteristic information. The profile analysis unit is used to generate the user's original payment profile based on the user's payment behavior and reminder contact behavior within a preset historical time window, and to determine the user's natural payment baseline based on natural payment samples. The collection strategy unit is used to determine the collection priority and reach matching comprehensive amount based on valid overdue bills, overdue characteristic information, user's original payment profile and user's natural payment baseline, generate differentiated collection strategies and form strategy intervention records; The feedback determination unit is used to determine the impact of the collection intervention based on the strategy intervention record, collection reach feedback information and payment feedback information, and to determine the credibility status of the feedback. The closed-loop update unit is used to update the user payment profile, save the effect sample of the collection strategy, or postpone the update based on the feedback credibility status, and form a collection closed loop status.

[0015] The technical effects and advantages of this invention are as follows: This invention links billing data, payment data, reminder contact data, and service interaction data through user identifiers, property identifiers, and billing cycles. It first identifies valid overdue bills and their characteristics, then creates a user's original payment profile based on historical payment and reminder contact behaviors. Compared to reminder methods that only sort by overdue amount or number of overdue days, this invention further filters natural payment samples and establishes a user's natural payment baseline. This allows the system to distinguish between a user's usual payment time and the time of payment received after being affected by reminder contacts. This provides a more reliable basis for judging reminder priority, contact matching, and differentiated reminder strategies, reducing misjudgments of user payment habits.

[0016] This invention, after executing a payment collection task, does not directly write the payment result into the user's payment profile. Instead, it determines the impact of the payment collection intervention by combining strategy intervention records, the timing of payment collection, the intensity of pre-payment outreach, and the pre-response status, and determines the reliability of the feedback accordingly. For payment results that reflect the user's original payment habits, the user's payment profile is updated; for payment results resulting from strong payment collection outreach, they are saved as samples of the payment collection strategy's effectiveness; for results with unclear bill attribution, fee disputes, abnormal contact information, or unclear contributions from multiple channels, updates are temporarily suspended. This reduces the contamination of long-term user profiles by strong payment collection and multi-channel overlapping outreach, improves the accuracy of subsequent payment collection strategy generation, outreach method selection, and payment collection intensity control, and forms a traceable payment collection closed loop. Attached Figure Description

[0017] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings; Figure 1 This is a flowchart illustrating the intelligent analysis method for precise payment collection based on user profiles according to the present invention. Figure 2 This is a schematic diagram illustrating the determination of the user's natural payment baseline according to the present invention; Figure 3 This is a schematic diagram illustrating the differentiated collection strategy and strategy intervention record generation of the present invention; Figure 4 This is a schematic diagram illustrating the impact of the payment reminder intervention on the determination of the reliability of the feedback in this invention. Figure 5 This is a schematic diagram of the structure of a precise payment collection intelligent analysis system based on user profiles according to the present invention. Detailed Implementation

[0018] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0019] Example 1: This example provides a precise intelligent analysis method for debt collection based on user profiles. This method is executed by an intelligent debt collection analysis platform. Figure 1 As shown, the method in this embodiment includes basic data acquisition, establishment of collection analysis transactions, determination of valid overdue bills, generation of user payment profiles, determination of user natural payment baseline, generation of differentiated collection strategies, formation of strategy intervention records, collection of collection outreach feedback and payment feedback, determination of the impact of collection intervention, judgment of the credibility of feedback, and formation of the collection closed loop status.

[0020] The intelligent payment collection analysis platform communicates with the property management fee collection system, user service platform, online payment platform, SMS platform, application messaging platform, outbound telephone system, and human customer service system to acquire billing data, payment data, payment collection reach data, and service interaction data. It also reliably updates user payment profiles based on payment collection feedback. Before acquiring basic data, the intelligent payment collection analysis platform reads the data access permissions and user contact information authorization status pre-configured by the property management company. For unauthorized or expired contact methods, no payment collection task is generated for the corresponding reach method. For historical payment records, historical payment collection records, and service interaction data used for statistical analysis, business associations are performed according to user identifiers, property identifiers, and payment cycles. Information unrelated to the current payment collection analysis transaction is not used as input for payment collection priority or reach matching.

[0021] First, basic data is acquired from the property management fee collection system, user service platform, online payment platform, and historical payment reminder database. Using user identifier, property identifier, and payment cycle as primary correlation fields, billing data, payment data, reminder outreach data, and service interaction data for the same user across different data sources are correlated to establish payment reminder analysis transactions. A payment reminder transaction number is generated for each transaction. The basic data includes pending bills, historical payment records, historical payment reminder records, outreach feedback records, payment channel usage records, and user service records. Pending bills include at least user identifier, property identifier, fee type, payment cycle, amount due, amount paid, due date, and bill status. Historical payment records include at least payment time, payment amount, payment channel, reimbursement bill, and payment entry source. Historical payment reminder records include at least reminder time, outreach method, number of outreach attempts, outreach result, and payment interval.

[0022] Before basic data enters the collection analysis process, the intelligent collection analysis platform performs field integrity checks on billing data, payment data, collection outreach data, and service interaction data. Data lacking user identifiers, property identifiers, or payment periods is not directly included in the current collection analysis process. Data lacking payment entry points, outreach results, or customer service notes is allowed to enter basic data storage, but is marked as incomplete evidence when determining the impact of collection intervention. For multiple billing records under the same user, property, and payment period, the billing record with the latest update time and outstanding status is prioritized. For payment records that cannot be mapped to a specific payment period, they are marked as pending verification of payment bill matching and are not used as the basis for updating the user's payment profile.

[0023] When establishing a payment collection analysis transaction, billing records for the same user, the same property, and the same billing cycle are processed for consistency. Bills with a status indicating they are settled, a payable amount less than or equal to the paid amount, or a current date not later than the due date are not considered valid overdue bills. Bills with a payable amount greater than the paid amount and a current date later than the due date are identified as valid overdue bills, and overdue characteristic information is generated based on the payable amount, paid amount, due date, billing cycle, and historical overdue records. This overdue characteristic information includes the overdue amount, number of overdue days, number of consecutive overdue cycles, cumulative number of overdue payments, and the interval between the most recent payment.

[0024] For users with valid overdue bills, the system reads their payment and outreach behavior within a preset historical time window to generate a raw user payment profile. This preset historical time window can be set to the most recent twelve months, the most recent six billing cycles, or determined based on the actual billing cycle corresponding to the billing type. The raw user payment profile includes the number of on-time payments, the number of overdue payments, the distribution of natural payment times, the number of payments made after reminders, historical SMS response records, historical phone call connection records, historical payment channel preferences, and historical payment intervals.

[0025] When generating the user's original payment profile, the user's natural payment baseline is further determined. This natural payment baseline represents the timeframe within which a user is likely to complete payment without strong collection intervention. To determine the user's natural payment baseline, natural payment samples are selected from historical payment records. These samples include payment records made within a billing cycle without outbound phone calls, customer service follow-ups, continuous outbound calls, or strong scripted reminders, as well as low-intervention payment records made only after receiving due date reminders, SMS reminders, or application message reminders. Payment records made after strong collection attempts are not directly used as the primary sample for the user's natural payment baseline. The user's natural payment baseline is obtained by statistically analyzing the number of days the payment time in the natural payment samples deviates from the due date. When the number of natural payment samples is insufficient, the user's natural payment baseline is marked as a low-reliability baseline.

[0026] After obtaining valid overdue bills, overdue payment characteristic information, the user's original payment profile, and the user's natural payment baseline, the overall priority and reach matching of collection efforts are determined, and a differentiated collection strategy is generated accordingly. In this embodiment, the user's natural payment baseline is not used as a regular user tag, but rather as a time reference for subsequently determining whether the current payment is subject to collection intervention. After the differentiated collection strategy is generated, the intelligent collection analysis platform simultaneously creates a strategy intervention record, enabling subsequent payment results to correspond with the current reach method, reach order, reach intensity, and prior response. Therefore, subsequent updates to the user's payment profile are not directly based on whether a payment has been made, but rather on whether the payment reflects the user's original payment habits.

[0027] The collection priority comprehensive quantity is used to indicate the order in which current overdue payment events should be processed in this round of collection tasks; the reach adaptation comprehensive quantity is used to indicate the degree of matching between a certain collection reach method and the current overdue user. The differentiated collection strategy includes collection reach methods, collection script templates, reach time windows, reach frequency, payment entry information, manual review identifiers, and strategy validity period. For low-priority collection matters, due date reminders, SMS reminders, or application message reminders are given priority; for medium-priority collection matters, a telephone reminder or delayed reminder can be arranged in addition to SMS reminders or application message reminders; for high-priority collection matters, tasks such as outbound calls, manual customer service follow-up, or paper notifications can be generated; for dispute verification matters, customer service verification tasks are generated first, and the regular collection reach process is only initiated after the dispute is resolved.

[0028] While generating differentiated collection strategies, a strategy intervention record is also generated. The strategy intervention record is used to record the degree of intervention that the collection strategy may have on the user's payment behavior, including at least the collection transaction number, contact method, contact order, contact time, number of contacts, intensity of the message, whether it includes human intervention, whether it includes a payment link, whether it is during the user's historical active payment period, and the contact level of the contact method in the current collection transaction.

[0029] The user's historical payment activity period is determined based on the user's payment time, payment entry click time, and application bill viewing time within a preset historical time window. The intelligent payment collection analysis platform counts the occurrence of the above behaviors by hourly or half-day periods, and determines the period when the occurrence frequency reaches a preset percentage threshold as the user's historical payment activity period; when the sample size is insufficient, the user's historical payment activity period is not generated, and it is marked as insufficient evidence of active period in the strategy intervention record.

[0030] The reach levels are categorized into low-intervention reach, medium-intervention reach, and high-intervention reach. Low-intervention reach includes due date reminders, SMS reminders, and app message reminders; medium-intervention reach includes repeated SMS reminders and outbound phone call reminders; and high-intervention reach includes follow-up by customer service representatives, continuous outbound calls, offline notifications, or reminders with explicit pressure to collect payments.

[0031] The strategy intervention records serve as the data basis for subsequently determining the intensity of pre-collection outreach, the status of pre-response, and the impact of collection intervention. When determining whether the collection result is suitable as a basis for updating the user's payment profile, the intelligent collection analysis platform reads the outreach order, outreach level, communication intensity, and whether there was human intervention from the strategy intervention records.

[0032] After the collection task is executed, collection reach feedback information and payment feedback information are collected. The collection reach feedback information includes SMS sending status, SMS delivery status, SMS click status, application message reading status, phone call connection status, phone communication result, customer service remarks, user reply content, payment link click status, and reason for failure to reach the customer. The payment feedback information includes payment time, payment amount, payment bill, payment channel, payment entry source, time interval between payment and the most recent reach, time interval between payment and the bill due date, and partial payment status. For cases where the corresponding bill for the payment cannot be determined, it is marked as a pending bill matching status and is not directly used to update the user's payment profile.

[0033] After obtaining feedback information on payment collection outreach and payment collection, the intelligent payment collection analysis platform does not directly update the user's payment profile based on whether the user has made a payment. Instead, it first reads the strategy intervention records and determines the impact of the payment collection intervention based on the timing of payment collection, the intensity of outreach before payment collection, and the status of the pre-payment response. The impact of the payment collection intervention indicates whether the current payment collection result reflects the user's original payment habits or whether it is primarily influenced by the differentiated payment collection strategy. The impact of the payment collection intervention includes low intervention impact status, medium intervention impact status, high intervention impact status, and attribution pending verification status.

[0034] After determining the impact of the payment collection intervention, the intelligent payment collection analysis platform combines payment collection feedback information, abnormal feedback information, and the user's original payment profile to determine the credibility status of the feedback. The credibility status of the feedback includes the profile updateable status, the strategy sample storage status, and the postponement update status, which are used to determine whether the current payment collection result will enter the user payment profile update process and be written into the payment collection strategy effect sample library, or be postponed for user payment profile updates.

[0035] After determining the credibility of the feedback, the updated user payment profile, saved strategy sample, or temporarily suspended update result are bound to the collection transaction number to form a closed-loop collection status. This closed-loop collection status includes at least the following states: payment received and user payment profile updated; payment received but strategy sample saved; payment received but update temporarily suspended; partial payment pending verification; non-payment under observation; non-payment transferred to high-priority status; and dispute investigation. When generating differentiated collection strategies, the user payment profile processed with the credible feedback status, the collection strategy effect sample library, and the closed-loop collection status are read to adjust the subsequent collection outreach methods and intensity.

[0036] In this embodiment, the payment result after the collection task is executed is not directly used as the basis for updating the user's payment profile. Instead, the impact of the collection intervention is first used to determine whether the payment has been affected by the current collection strategy. Then, the credibility status of the feedback determines whether to update the profile, save the strategy sample, or postpone the update. This can reduce the reverse pollution of the user's payment profile caused by strong collection outreach, multi-channel overlapping outreach, and abnormal bill attribution.

[0037] Example 2: This example further explains the method for determining the user's natural payment baseline, based on Example 1; for example... Figure 2 As shown.

[0038] When generating a user's original payment profile, a sample of natural payments is selected from historical payment records. This natural payment sample includes payment records made within a billing cycle during which no outbound calls, customer service follow-ups, continuous outbound calls, or aggressive reminders occurred, as well as low-intervention payment records made only after receiving due date reminders, SMS reminders, or application message reminders. Payment records made after aggressive collection efforts are not directly used as the primary sample for the user's natural payment baseline.

[0039] When determining the baseline for users' natural payment, the number of days the payment time is offset from the due date is used as the analysis object. The offset number of days is the interval between the payment time and the due date; a payment time earlier than the due date is recorded as a negative offset, and a payment time later than the due date is recorded as a positive offset. By statistically analyzing the offset number of days in the natural payment sample, the common payment time range for users under normal circumstances is obtained.

[0040] When the number of natural payment samples reaches a preset sample size threshold, the median value of the number of offset days in the natural payment samples is taken as the natural payment benchmark, and the allowable range of natural payments is determined based on the dispersion of the offset days. The preset sample size threshold is three valid payment samples; when natural payment samples are statistically analyzed by payment cycle, the preset sample size threshold is three payment cycles. The allowable range of natural payments is determined based on the concentrated distribution interval of the offset days. In one embodiment, the allowable range is one to three days before and after the natural payment benchmark; in another embodiment, the allowable range is the interval between the upper and lower quantile values ​​of the offset days in the natural payment samples.

[0041] When the number of natural payment samples is insufficient, the user's original payment profile is retained, but the user's natural payment baseline is marked as a low-confidence baseline. For low-confidence baselines, when determining the impact of collection interventions and the credibility of feedback, the determination of whether the current payment represents the user's original payment habits is not based solely on this baseline. Instead, the judgment is made in combination with the intensity of contact before payment, the status of the pre-payment response, abnormal feedback information, and historical collection response.

[0042] Under a low-reliability baseline, the intelligent debt collection analysis platform processes the repayment result according to a cautious update principle. If the repayment bill matches the current valid outstanding bill, and only due reminders, SMS reminders, or app message reminders occurred before the repayment, and there were no disputes, complaints, abnormal contact information, or unclear contributions from multiple channels, the feedback reliability status can be determined as an updatable profile status, and the user's payment profile can be adjusted with a low-amplitude update method. If the repayment occurs after outbound calls, customer service follow-ups, continuous outbound calls, or offline notifications, or if the repayment bill cannot clearly match the current valid outstanding bill, the repayment result will not be used as the basis for updating the user's payment profile.

[0043] This embodiment distinguishes between natural payment samples and payment records formed after being contacted by forced collection, so that the user's natural payment baseline can reflect the payment time pattern of the user when not subject to forced collection intervention, providing a comparable time benchmark for subsequent judgment of the impact of collection intervention.

[0044] Example 3: This example, based on Example 1 or Example 2, further explains the generation methods of the collection priority summary, the reach matching summary, and the strategy intervention record; such as... Figure 3 As shown.

[0045] After obtaining valid overdue bills, overdue characteristic information, the user's original payment profile, and the user's natural payment baseline, a comprehensive collection priority is determined. This comprehensive collection priority is determined based on the level of overdue amount, the level of overdue days, the level of consecutive overdue periods, the level of historical collection response, and the level of impact of service disputes. Each input factor can be converted to the same level range, which can be set to 0 to 3 or 0 to 5 levels. A higher level indicates a stronger effect of the corresponding factor on increasing the collection priority.

[0046] In practice, all input factors can be uniformly converted to levels 0 to 3. The level of outstanding payment amount is determined based on the distribution of the outstanding amount of the current valid overdue bill within the community, the same fee type, or the same billing cycle; for example, an outstanding amount lower than the median of similar bills is recorded as level 0; reaching the median but not the upper quartile is recorded as level 1; reaching the upper quartile but not exceeding the preset large outstanding amount is recorded as level 2; and exceeding the preset large outstanding amount is recorded as level 3. The preset large outstanding amount can be determined according to the property management fee standard, for example, set as three times the amount due in a single billing cycle, or set as 3,000 to 5,000 yuan. The level of overdue days is determined based on the number of days past the due date; for example, overdue for no more than three days is recorded as level 0; overdue for more than three days but no more than seven days is recorded as level 1; overdue for more than seven days but no more than thirty days is recorded as level 2; and overdue for more than thirty days is recorded as level 3. The continuous arrears cycle level is determined based on the number of consecutive periods in which the same user has valid arrears in adjacent billing cycles. For example, no consecutive arrears are recorded as level 0; two consecutive billing cycles of arrears are recorded as level 1; three consecutive billing cycles of arrears are recorded as level 2; and four or more consecutive billing cycles of arrears are recorded as level 3. The historical collection response level is determined based on the number of times the user made a payment after being contacted in historical collection tasks, the average payment interval, and the number of times there was no response. For example, if the user made a payment within one day after being contacted with low intervention in the last three collection cycles, it is recorded as level 0; if there was one instance of no payment after being contacted in the last three collection cycles, it is recorded as level 1; if there were two instances of no payment after being contacted or an average payment interval exceeding seven days, it is recorded as level 2; and if there were three consecutive instances of no payment after being contacted or an average payment interval exceeding fifteen days, it is recorded as level 3. The impact level of a service dispute is determined based on unclosed complaints, unclosed repair work orders, fee dispute records, and manually verified items. When there are unclosed fee disputes, unclosed complaints, or unclosed manually verified items, the impact level of the service dispute will be determined as Level 3, and the corresponding collection analysis will be transferred to dispute verification and processing. This will not be used as a basis for increasing collection intensity.

[0047] In one specific implementation, the overall collection priority is determined by the level of overdue amount, the level of overdue days, the level of consecutive overdue periods, and the level of historical collection response. Without a dedicated weight, the overall collection priority is determined by summing these four levels; with a dedicated weight, the four levels are weighted according to preset weights. These preset weights are pre-set by the property management company based on the fee type, collection period, and historical payment data, and remain consistent within the same collection batch. When using a weighted method, the weights of the overdue days level and the consecutive overdue period level are higher than the overdue amount level, and the historical collection response level is used to correct the user's response to previous collection attempts. Collection priorities reaching the preset high-priority threshold are classified as high-priority collection transactions; reaching the preset medium-priority threshold but not the preset high-priority threshold are classified as medium-priority collection transactions; and falling below the preset medium-priority threshold are classified as low-priority collection transactions.

[0048] Within the same collection batch, the input factors, level ranges, preset weights, preset medium-priority thresholds, and preset high-priority thresholds of the collection priority are kept consistent and bound to the collection batch number for storage. If the collection loop status changes subsequently, the system does not overwrite the collection priority already formed in that collection batch, but instead redetermines it in the next collection batch based on the updated user payment profile and collection loop status.

[0049] After determining the priority of payment collection, a comprehensive reach matching score is determined based on the user's original payment profile, contact authorization status, historical contact records, payment channel usage records, and the user's natural payment baseline. This comprehensive reach matching score represents the degree of matching between a particular payment collection method and the current user with outstanding payments. The comprehensive reach matching score is determined separately for SMS reminders, app message reminders, outbound phone calls, customer service follow-ups, and paper notifications. Contact methods that do not meet the requirements of contact authorization status, invalid contact information, or have reached the contact frequency limit are excluded or downgraded.

[0050] In one specific implementation, the overall reach matching quantity is generated separately according to different debt collection reach methods. Each debt collection reach method corresponds to one overall reach matching quantity, and a single overall quantity is not used to replace the matching judgment of all reach methods. The debt collection intelligent analysis platform determines the overall reach matching quantity based on the historical reach success rate, post-reach payment interval, payment gateway response status, user's commonly used payment channels, contact information validity status, reach frequency restriction status, and user's natural payment baseline for the reach method, and classifies the corresponding reach method into high-fit reach method, medium-fit reach method, low-fit reach method, or unavailable reach method.

[0051] The higher the success rate of outreach, the shorter the repayment interval after outreach, and the more times the payment gateway responds, the higher the overall outreach compatibility. If the contact information is invalid, unauthorized, or the outreach frequency reaches the preset frequency limit, the corresponding outreach method will be downgraded to a low-compatibility outreach method or an unavailable outreach method. Unavailable outreach methods will not be included in the current collection outreach order; low-compatibility outreach methods will only be used as backup methods when other outreach methods are unavailable or require manual review.

[0052] In practice, the intelligent debt collection analysis platform determines the overall reach suitability of SMS reminders, app message reminders, outbound phone calls, customer service follow-ups, and paper notifications. For each debt collection method, it reads the number of successful contacts, the number of failed contacts, the payment interval after contact, the number of clicks on the payment portal, the authorization status of the contact information, the validity status of the contact information, and the frequency restriction status of the contact within a preset historical time window. Taking SMS reminders as an example, if the SMS delivery rate reaches 80% or more in the last six billing cycles, and the payment is completed more than twice within 24 hours after clicking the payment portal in the SMS, the SMS reminder is identified as a high-suitability reach method; if the SMS delivery rate is 50% or more but less than 80%, or the payment is completed only once within 24 hours after clicking the payment portal in the SMS, the SMS reminder is identified as a medium-suitability reach method; if the SMS delivery rate is less than 50%, or three consecutive SMS messages fail to be sent, the SMS reminder is identified as a low-suitability reach method or an unusable reach method. Taking outbound phone calls as an example, if the call connection rate reaches 60% or higher within the last six billing cycles, and payment is completed at least twice within three days after the call reminder, the outbound phone call will be classified as a high-adaptability reach method; if the call connection rate is 30% or higher but less than 60%, the outbound phone call will be classified as a medium-adaptability reach method; if the call connection rate is less than 30%, or if there are three consecutive unsuccessful calls, the outbound phone call will be classified as a low-adaptability reach method. Contact methods that are unauthorized, invalid, or have reached the preset frequency limit will not be considered as the current priority reach method. The preset frequency limit can be set to no more than three SMS reminders, no more than three application message reminders, and no more than two outbound phone calls within the same billing cycle.

[0053] For each collection outreach method, the intelligent collection analysis platform separately stores the input basis and matching results of the outreach adaptation comprehensive volume. If a collection outreach method is determined to be an unavailable outreach method due to unauthorized contact information, invalid contact information, or reaching the preset frequency limit, then this collection outreach method will not be included in the current collection outreach order, nor will it be considered a valid outreach node in subsequent payment attribution.

[0054] After generating a comprehensive volume of reach matching for each payment collection method, the intelligent payment collection analysis platform combines the user's natural payment baseline and abnormal feedback records to determine the final contact order. If the user's natural payment baseline shows that they typically complete payments within the allowed natural payment range, or exceed the allowed range but not exceed a preset minor deviation period, then due date reminders, SMS reminders, or app message reminders will be prioritized; the preset minor deviation period can be set to one to two days. If the user has a record of payment disputes, unclosed complaints, or matters requiring manual verification, then a manual customer service follow-up task for payment verification, bill explanation, or dispute resolution will be generated first, without directly entering the forced payment collection process. If the user's contact information is invalid, or the number of consecutive failed contacts reaches a preset number of failures, a contact information review task will be generated; the preset number of failures can be set to three.

[0055] Differentiated collection strategies are generated based on collection priority results and overall reach matching volume. These strategies include collection reach methods, collection script templates, reach time windows, reach frequency, payment entry information, manual review indicators, and strategy validity periods. For low-priority collection transactions, due date reminders, SMS reminders, or application message reminders are generated first, using reminder-style scripts. For medium-priority collection transactions, in addition to SMS or application message reminders, a phone reminder is arranged or a delayed reminder is given based on the overall reach matching volume. For high-priority collection transactions, outbound phone calls, human customer service follow-ups, or paper notification tasks can be generated, but the reach intensity is still controlled by considering reach frequency limits and abnormal feedback records. For dispute verification transactions, customer service verification tasks are generated first, and the regular collection reach process only begins after the dispute is resolved.

[0056] While generating differentiated payment collection strategies, a strategy intervention record is also generated. This record includes at least the payment collection transaction number, contact method, contact order, contact time, number of contacts, message intensity, whether it includes human intervention, whether it includes a payment link, whether it occurs during a period of active user payment history, and the contact level of that method within the current payment collection transaction. Message intensity can be categorized based on whether the message template only includes payment reminders, indicates overdue status, prompts for human follow-up, or suggests subsequent handling measures. Messages only including payment due reminders, bill amount, and payment entry prompts are classified as low intensity; messages including overdue status prompts, reminders again, or payment deadline prompts are classified as medium intensity; and messages including prompts for human follow-up, offline notifications, or subsequent handling measures are classified as high intensity. If a single contact record contains multiple intensity levels, the record with the highest intensity level is used.

[0057] In this embodiment, the strategy intervention record is used to determine whether the payment was affected by the collection strategy after the payment has been made. This record is not a regular execution log, but rather the data basis for subsequently determining the intensity of outreach before payment, the status of the pre-payment response, and the impact of the collection intervention.

[0058] Example 4: This example further explains the method for determining the impact of the payment reminder intervention and the reliability of the feedback, based on any of the above examples.

[0059] After the collection task is executed, the system receives execution results from the SMS platform, application messaging platform, outbound call system, and human customer service system, and writes them into the corresponding collection analysis transaction according to the collection transaction number. The execution results include SMS sending status, SMS delivery status, SMS click status, application message reading status, call connection status, call communication result, human customer service remarks, user reply content, payment link click status, and reason for failure to reach the customer. These execution results are then organized in chronological order of reach to form collection reach feedback information.

[0060] Within a preset observation window after the collection task is executed, payment results from the online payment platform and property management fee system are read, and the payment results are matched with the collection task number, user identifier, house identifier, fee type, and fee cycle. The preset observation window can be set to one to seven days after the collection task is executed; for annual property fees or high-amount comprehensive bills, it can also be set to seven to fifteen days. If the user completes payment within the preset observation window, a payment feedback message is generated; if the user only completes partial payment, a partial payment status is generated, and the remaining outstanding amount is recorded; if no payment result is detected, a non-payment status is generated. For cases where the corresponding bill for payment cannot be determined, it is marked as a pending bill matching status.

[0061] After obtaining feedback information on payment reminders and payment collection, instead of directly updating the user's payment profile based on the payment collection results, the impact of payment reminder intervention is determined. For example... Figure 4 As shown, the intelligent debt collection analysis platform determines the impact of debt collection intervention based on the timing of payment, the intensity of pre-payment outreach, and the pre-response status. It further combines payment feedback information, abnormal feedback information, and the user's original payment profile to determine the credibility of the feedback, thus categorizing the current payment result as either profile updatable, strategy sample preserved, or temporarily deferred. The impact of debt collection intervention is determined based on three evidentiary states: the timing of payment, the intensity of pre-payment outreach, and the pre-response status.

[0062] The repayment timing relationship refers to the relationship between the current repayment time and the user's natural payment baseline. The number of days the current repayment time deviates from the payment due date is compared to the user's natural payment baseline. If the current repayment time falls within the normal payment time range corresponding to the user's natural payment baseline, or exceeds this range but does not exceed a preset slight deviation period, it is determined to be close to the natural payment baseline. The preset slight deviation period can be set to one to two days. If the current repayment time exceeds the normal payment time range corresponding to the user's natural payment baseline, and the excess period reaches a preset significant deviation period, it is determined to be deviating from the natural payment baseline. The preset significant deviation period can be set to three to five days. If the user's natural payment baseline is marked as a low-confidence baseline, the repayment timing relationship is marked as baseline insufficiency.

[0063] The term "pre-payment contact intensity" refers to the actual intensity of collection contact experienced by the user before payment is completed. Only contact actions occurring before the payment due date are counted. If the user only receives due reminders, SMS reminders, or app message reminders before payment, and there are no outbound calls, customer service follow-ups, continuous outbound calls, or offline notifications, it is determined to be low-intensity contact. If the user receives repeated SMS reminders, secondary app message reminders, one outbound call, or one customer service reminder before payment, but there are no continuous outbound calls, offline notifications, or aggressive collection tactics, it is determined to be medium-intensity contact. If the user receives continuous outbound calls, customer service follow-ups, offline notifications, or if the contact tactics contain clear follow-up measures, deadline reminders, or other collection pressure content before payment, it is determined to be high-intensity contact.

[0064] The pre-emptive response status refers to whether the user has already effectively responded to the low-intervention outreach before the strong outreach occurs. Strong outreach includes outbound phone calls, customer service follow-ups, continuous outbound calls, and offline notifications; effective responses include clicking on SMS messages, reading application messages, accessing payment links, opening payment pages, users proactively inquiring about payment entry points, or users accessing bill details pages within the application. If the above-mentioned effective responses exist before the strong outreach occurs, it is determined that a pre-emptive response exists; if no effective response is detected before the strong outreach occurs, and the user completes payment within a preset short time interval after the strong outreach, it is determined that there is no pre-emptive response and payment is made after the strong outreach; if multiple outreach methods occur consecutively in the same payment collection transaction, and multiple outreach nodes have response records, but it cannot be determined which outreach method primarily facilitated the final payment, it is determined that the pre-emptive response attribution is unclear. The preset short time interval can be set to two hours to twenty-four hours.

[0065] When determining the impact of collection intervention based on the timing of payment, the intensity of pre-payment outreach, and the status of pre-payment response, the intelligent collection analysis platform follows a sequence of first excluding anomalies, then determining high intervention levels, and finally low and medium intervention levels. If there are pending verification statuses for payment bills, fee disputes, complaint feedback, abnormal contact information, or unclear contributions from multiple outreach channels, the platform prioritizes determining the status as pending verification. If none of the above situations exist, the platform then determines whether the current payment falls under a high intervention impact status. If the high intervention impact status is not met, the platform further considers the user's natural payment baseline and the intensity of pre-payment outreach to determine whether it falls under a low or medium intervention impact status.

[0066] After determining the above three evidentiary states, the impact of collection intervention will be categorized as low-intervention impact, medium-intervention impact, high-intervention impact, or attribution pending verification. If the repayment time is close to the natural payment baseline, and the pre-repayment outreach intensity is low, the collection intervention impact will be classified as low. If the repayment time is close to the user's natural payment baseline, or although it exceeds the normal payment time range corresponding to the user's natural payment baseline but does not reach the preset significant deviation number of days, and the pre-repayment outreach intensity is medium, the collection intervention impact will be classified as medium. If the repayment time deviates from the natural payment baseline, the pre-repayment outreach intensity is high, and there was no prior response before the strong outreach occurred, the collection intervention impact will be classified as high. If the user had a prior response before the high-intensity outreach, or the contributions of multiple outreach channels in the same collection transaction are unclear, making it impossible to determine which outreach method primarily facilitated the final repayment, the collection intervention impact will be classified as attribution pending verification.

[0067] For non-repayment status, no collection intervention impact value indicating the source of repayment is generated, but the contact intensity before repayment, the reason for contact failure, and the user's response content are retained, and this result is used as input for subsequent feedback credibility determination. For partially repaid status, it is first determined whether the partial repayment can match the current valid outstanding bill; if it can match, the collection intervention impact value is determined according to the actual repayment time; if it cannot match, it is transferred to the attribution pending verification status or the repayment bill matching pending verification status.

[0068] After determining the impact of the collection intervention, the credibility of the feedback is determined by combining payment feedback information, abnormal feedback information, and the user's original payment profile. Abnormal feedback information includes fee disputes, unclosed complaints, discrepancies in bill amounts, inconsistent property identification, inconsistent billing cycles, abnormal contact information, users explicitly refusing to acknowledge the bill, users reporting that they have paid through other channels but have not completed bill reconciliation, payment records indicating a risk of duplicate reconciliation, and situations where customer service notes indicate a need for review. When abnormal feedback information exists, the intelligent collection analysis platform does not directly update the user's payment profile based on the current payment result. Instead, it determines whether to maintain the strategy sample or postpone updates based on the type of abnormal feedback information. When determining the credibility of the feedback, the intelligent collection analysis platform does not use the payment result itself as the sole criterion. Instead, it considers the impact of the collection intervention, bill matching results, abnormal feedback information, and the user's historical payment patterns to determine whether the payment result can be used as the basis for updating the user's payment profile.

[0069] When the impact of payment reminder intervention is low, and the payment feedback information shows that the user completed payment within the preset observation time window, and the payment bill matches the current valid outstanding bill, and there is no abnormal feedback information, the feedback credibility status is determined to be an updatable profile status. In the updatable profile status, the user's payment profile is updated based on this payment result. If the user completes payment promptly after low-intensity contact, the credibility of the reminder response tag is increased; if the user completes payment close to their natural payment baseline across multiple billing cycles, the credibility of the payment stability tag is increased; if the user fails to make payment multiple times after low-intensity contact and there is no abnormal feedback information, the credibility of the overdue habit tag is increased.

[0070] In practice, the user payment profile can include the credibility of reminder response tags, the credibility of payment stability tags, the credibility of overdue payment habit tags, and the credibility of payment channel preference tags. Each tag's credibility can be represented by a value between 0 and 1; the higher the value, the stronger the reference value of the corresponding tag for subsequent differentiated collection strategies. The intelligent collection analysis platform determines the initial value of each tag's credibility based on the number of valid samples within historical payment periods. For example, if the number of valid samples is less than three payment periods, the initial credibility value of the corresponding tag should not exceed 0.3; if the number of valid samples is three to six payment periods, the initial credibility value can be set to 0.3 to 0.6; if the number of valid samples is more than six payment periods and the feedback results are stable, the initial credibility value can be set to 0.6 to 0.8. The user payment profile is only adjusted when the feedback credibility status is in an updatable state. For profiles created under low-intervention impact conditions that are ready for updating, the single-time label credibility adjustment range can be set to 0.05 to 0.15; for profiles created under medium-intervention impact conditions that are ready for updating, the single-time label credibility adjustment range can be set to 0.01 to 0.05. For strategy sample preservation and delayed update states, the label credibility in the user payment profile is not adjusted. Adjusted label credibility values ​​below 0 are treated as 0; values ​​above 1 are treated as 1.

[0071] When the impact of the payment reminder intervention is at the medium level, the credibility of the feedback is further determined by combining the user's original payment profile and abnormal feedback information. If the current payment time exceeds the normal payment time range corresponding to the user's natural payment baseline, but the number of days exceeding the limit does not reach the preset significant deviation number, and the current payment amount matches the current valid outstanding bill, the payment channel belongs to the user's historically frequently used payment channel, and there are no fee disputes, complaints, unclear bill matching, or unclear contributions from multiple channels, the credibility of the feedback is determined to be a profile updateable state, and the user's payment profile is updated at a low level. The preset significant deviation number can be set to three to five days; the low level update refers to a single tag credibility adjustment range of 0.01 to 0.05. If the current payment mainly occurs after repeated reminders, one outbound phone call, or one manual reminder, and the payment time, payment channel, or user feedback content cannot correspond to the user's historical payment patterns, the credibility of the feedback is determined to be a strategy sample preservation state.

[0072] When the impact of a payment collection intervention is at a high level, and the user completes payment within a preset short time interval following continuous outbound calls, customer service follow-ups, offline notifications, or aggressive payment reminders, the feedback credibility status is determined as the strategy sample storage status. The preset short time interval can be set to two to twenty-four hours. In this status, the payment collection method, contact order, contact intensity, payment time, payment amount, user response content, customer service remarks, and abnormal feedback information are written into the payment collection strategy effect sample library, but the payment result is not directly written into the user payment profile.

[0073] When writing to the collection strategy effectiveness sample library, the intelligent collection analysis platform simultaneously saves the collection transaction number, valid overdue bill identifier, collection priority comprehensive value, reach adaptation comprehensive value for each collection outreach method, strategy intervention record, collection intervention impact value, and feedback credibility status. This collection strategy effectiveness sample library is used for subsequent evaluation of the execution effect of differentiated collection strategies and is not used as a sample source for directly updating user payment profiles; only when subsequent manual review confirms that the current payment reflects the user's original payment habits will the sample be allowed to be transferred to the user payment profile update process.

[0074] When the impact of collection intervention is in an attribution pending verification state, or when there are situations such as pending verification of payment bill matching, inability to attribute some payments, fee disputes, complaint feedback, abnormal contact information, or unclear contributions from multiple channels, the feedback credibility status will be determined as a temporarily suspended update state. For users whose natural payment baseline is a low credibility baseline, the user payment profile will not be updated solely based on the payment timing; only when the payment bill matching is clear, only low-intensity outreach occurred before the payment, and there is no abnormal feedback information, will a low-amplitude profile update be performed in conjunction with historical collection response data. In the temporarily suspended update state, the user's original payment profile remains unchanged, and only the current collection outreach feedback information, payment feedback information, abnormal feedback information, and the impact of collection intervention are written to the collection analysis record database.

[0075] For accounts in a "temporarily suspended update" state, the intelligent payment collection analysis platform generates a "pending verification" flag and records the reason for the suspension. The reasons for the suspension include pending verification of payment bill matching, inability to attribute some payments, unresolved payment disputes, unclosed complaint feedback, abnormal contact information, unclear contributions from multiple channels, or a low-credibility baseline for the user's natural payment. Before the "pending verification" flag is removed, this payment result will not be included in the updates to the credibility of the reminder response tag, payment stability tag, overdue habit tag, and payment channel preference tag in the user's payment profile. After the "pending verification" flag is removed, the system will, based on the review results, change the payment result to a profile updateable state, a strategy sample preservation state, or continue to maintain the "temporarily suspended update" state.

[0076] After determining the credibility of the feedback, the updated user payment profile, the saved strategy sample, or the postponed update result are bound to the collection transaction number to form a closed-loop collection status. This closed-loop collection status includes at least the following states: payment received and user payment profile updated; payment received but strategy sample saved; payment received but update postponed; partial payment pending verification; non-payment under observation; non-payment transferred to high-priority status; and dispute investigation. When generating differentiated collection strategies, the user payment profile and the closed-loop collection status after the feedback credibility determination are read first.

[0077] After the above processing, the payment results will not be directly written into the user's payment profile. Instead, they will first be assessed based on the impact of the collection intervention and the reliability of the feedback. Payment results that are closer to the user's original payment habits will be used to update the user's payment profile. Payment results resulting from aggressive collection efforts will be saved as a sample of the strategy's effectiveness. For results with unclear contributions from multiple channels or abnormal bill attribution, updates to the user's payment profile will be temporarily suspended, thereby reducing the negative contamination of the user's payment profile by collection actions.

[0078] Example 5: This example provides a precise payment collection intelligent analysis system based on user profiles, used to execute the precise payment collection intelligent analysis method based on user profiles described in any of the above examples. The system is deployed in a payment collection intelligent analysis platform and communicates with a property management fee collection system, a user service platform, an online payment platform, an SMS platform, an application messaging platform, a telephone outbound calling system, and a human customer service system. The system includes a data processing unit, a profile analysis unit, a payment collection strategy unit, a feedback judgment unit, and a closed-loop update unit; such as... Figure 5 As shown.

[0079] The data processing unit is used to acquire pending bills, historical payment records, historical reminder records, contact feedback records, payment channel usage records, and user service records. Using user identifier, property identifier, and billing cycle as primary correlation fields, it associates bill data, payment data, reminder contact data, and service interaction data for the same user across different data sources, establishing reminder analysis transactions and generating reminder transaction numbers. The data processing unit is also used to perform consistency processing on bill records for the same user, property, and billing cycle, identifying valid overdue bills and generating overdue feature information.

[0080] The profiling analysis unit is used to generate the user's original payment profile based on the user's payment behavior and reminder contact behavior within a preset historical time window, and to filter natural payment samples from historical payment records. The user's natural payment baseline is determined based on the number of days the payment time in the natural payment sample is offset from the due date. When the number of natural payment samples is insufficient, the user's natural payment baseline is marked as a low-confidence baseline, and the low-confidence baseline status is provided to the reminder strategy unit and the feedback judgment unit for invocation.

[0081] The collection strategy unit is used to determine the comprehensive collection priority and reach matching comprehensive values ​​based on valid overdue bills, overdue characteristic information, the user's original payment profile, and the user's natural payment baseline, and to generate differentiated collection strategies. The collection strategy unit is also used to generate strategy intervention records when generating differentiated collection strategies. These records include at least the collection transaction number, reach method, reach order, reach time, number of reach attempts, intensity of the communication, whether it includes human intervention, whether it includes a payment link, whether it falls within the user's historical active payment period, and the reach level.

[0082] The feedback determination unit is used to collect collection reach feedback information and payment feedback information after the collection task is executed, and to read strategy intervention records. It determines the impact of collection intervention based on the timing of payment, the intensity of reach before payment, and the pre-response status. The feedback determination unit is also used to determine the reliability status of feedback by combining the impact of collection intervention, payment feedback information, abnormal feedback information, and the user's original payment profile. The reliability status includes a profile updateable state, a strategy sample preservation state, and a postponed update state. The reliability status is used to distinguish whether the current payment result can be used as the basis for updating the user's payment profile.

[0083] The closed-loop update unit is used to perform corresponding processing based on the feedback credibility status: when the feedback credibility status is "profile updateable," the user's payment profile is updated; when the feedback credibility status is "strategy sample saving," the current collection method, contact order, contact intensity, repayment time, repayment amount, and user feedback content are written into the collection strategy effect sample library, but the user's payment profile is not directly updated; when the feedback credibility status is "update postponed," the user's original payment profile remains unchanged. The closed-loop update unit is also used to bind the user payment profile update result, strategy sample saving result, or postponed update result with the collection transaction number, forming a collection closed-loop status, which can be called upon when generating differentiated collection strategies later. The closed-loop update unit is also used to save records of tag credibility changes before and after the user payment profile update, records of writing to the collection strategy effect sample library, and records of postponed update reasons, and bind these records with the collection transaction number.

[0084] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0085] In the several embodiments provided in this application, it should be understood that the disclosed systems, apparatuses, and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0086] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0087] In addition, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit.

[0088] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A precise intelligent analysis method for payment collection based on user profiles, characterized in that, Includes the following steps: Acquire basic data, establish a collection analysis process, and identify valid overdue bills and their corresponding overdue characteristic information; For users with valid overdue bills, an original payment profile is generated based on historical payment behavior and reminder contact behavior. Natural payment samples are then selected from historical payment records, and the user's natural payment baseline is determined based on the natural payment samples. Based on the valid overdue bills, overdue feature information, user's original payment profile, and user's natural payment baseline, the collection priority comprehensive quantity is determined, and the corresponding reach adaptation comprehensive quantity is determined for different collection reach methods. Based on the collection priority comprehensive quantity and the reach adaptation comprehensive quantity corresponding to each collection reach method, a differentiated collection strategy is generated, and a strategy intervention record is formed at the same time. After the collection task is executed, the collection reach feedback information and payment feedback information are obtained. Based on the strategy intervention record, the relationship between payment time, the intensity of reach before payment and the pre-repayment response status, the impact of the collection intervention is determined. The credibility status of the feedback is determined based on the impact of the payment reminder intervention, and the user payment profile is updated, the effect sample of the payment reminder strategy is saved or the update is temporarily suspended according to the credibility status of the feedback, thus forming a closed loop of payment reminder.

2. The intelligent analysis method for precise payment collection based on user profiles according to claim 1, characterized in that: The basic data includes pending bills, historical payment records, and historical reminder records, and includes at least one of the following: contact feedback records, payment channel usage records, and user service records. When establishing a reminder analysis transaction, user identifier, house identifier, and payment cycle are used as the main correlation fields to associate the bill data, payment data, reminder contact data, and service interaction data of the same user in different data sources, and generate a reminder transaction number.

3. The intelligent analysis method for precise payment collection based on user profiles according to claim 2, characterized in that: When determining valid overdue bills, consistent processing is performed on bill records for the same user, the same property, and the same billing cycle. Bills with an amount due greater than the amount paid and a current date later than the due date are determined to be valid overdue bills. Bills that do not meet the above conditions are considered invalid overdue bills and are not included in the current collection process. The overdue characteristic information includes at least one of the following: overdue amount, number of overdue days, number of consecutive overdue cycles, cumulative number of overdue payments, and the interval between the most recent payment.

4. The intelligent analysis method for precise payment collection based on user profiles according to claim 1, characterized in that: The original user payment profile is generated based on the user's payment behavior and reminder contact behavior within a preset historical time window. The original user payment profile includes at least one of the following: number of on-time payments, number of overdue payments, distribution of natural payment time, number of payments after reminders, historical SMS response status, historical phone call connection status, historical payment channel preference, and historical payment interval.

5. The intelligent analysis method for precise payment collection based on user profiles according to claim 1, characterized in that: When selecting natural payment samples, payment records formed within the billing cycle that did not involve outbound phone calls, customer service follow-ups, continuous outbound calls, or strong scripted reminders are considered natural payment samples. Low-intervention payment records formed only after receiving due date reminders, SMS reminders, or application message reminders are also considered natural payment samples. Payment records formed after strong collection efforts have occurred are excluded from the user's natural payment baseline sample.

6. The intelligent analysis method for precise payment collection based on user profiles according to claim 5, characterized in that: When determining the user's natural payment baseline, the number of days the payment time is offset from the due date is used as the analysis object; When the number of natural payment samples reaches a preset sample number threshold, the user's natural payment baseline is determined based on the median value of the number of offset days in the natural payment samples and the allowable range of natural payment. The allowable range of natural payment is determined based on the dispersion of the number of offset days, or by the interval between the upper and lower quantile values ​​of the number of offset days. When the number of natural payment samples is less than the preset sample number threshold, the user's natural payment baseline is marked as a low-confidence baseline.

7. The intelligent analysis method for precise payment collection based on user profiles according to claim 1, characterized in that: When determining the overall priority of collection, it is determined based on at least some of the following: level of overdue amount, level of overdue days, level of consecutive overdue period, level of historical collection response, and level of impact of service disputes. When there are unclosed fee disputes, unclosed complaints, or unclosed matters under manual verification, the level of impact of service disputes is set to the highest preset level, and the corresponding collection analysis matter is transferred to the dispute verification and processing process, and the regular collection intensity escalation operation is suspended.

8. The intelligent analysis method for precise payment collection based on user profiles according to claim 1, characterized in that: When determining the overall reach matching volume, corresponding overall reach matching volumes are generated for different debt collection reach methods. Based on the historical reach success rate, post-reach payment interval, payment gateway response status, user's commonly used payment channels, contact information validity status, reach frequency limit status, and user's natural payment baseline, the corresponding debt collection reach methods are classified into high-fit reach methods, medium-fit reach methods, low-fit reach methods, or unavailable reach methods. Among them, unavailable reach methods are not included in the current debt collection reach order, and low-fit reach methods are only used as alternative methods when other reach methods are unavailable or require manual review.

9. The intelligent analysis method for precise payment collection based on user profiles according to claim 1, characterized in that: The strategy intervention record includes at least the collection transaction number, contact method, contact order, contact time, number of contacts, intensity of the communication, whether it includes human intervention, whether it includes a payment link, whether it falls within the user's historical active payment period, and the contact level. When determining the impact of the collection intervention, the timing of the payment is determined based on the relationship between the current payment time and the user's natural payment baseline; the pre-payment contact intensity is determined based on the actual collection contact intensity experienced by the user before completing the payment; and the pre-response status is determined based on whether there was an effective response before the strong contact occurred. The strong contact includes outbound phone calls. The following measures will be taken: customer service follow-up, continuous outbound calls, or offline notifications. If the current payment is close to the user's natural payment baseline and the outreach intensity before the payment is low, it will be classified as a low-intervention-impact state. If the current payment deviates from the user's natural payment baseline, the outreach intensity before the payment is high, and there was no effective response behavior before the strong outreach occurred, it will be classified as a high-intervention-impact state, and the corresponding payment result will be saved as a sample of the effectiveness of the collection strategy, and will not be directly used to update the user's payment profile. If the contribution of multiple channels in the same collection transaction is unclear, it will be classified as an attribution pending verification state and the update will be temporarily suspended.

10. A precise payment collection intelligent analysis system based on user profiles, characterized in that, The system is deployed on a payment collection intelligent analysis platform and is connected to at least one of a property management fee collection system, a user service platform, an online payment platform, an SMS platform, an application messaging platform, a telephone outbound calling system, and a human customer service system. It is used to execute the precise payment collection intelligent analysis method based on user profiles as described in any one of claims 1 to 9. The system includes: The data processing unit is used to acquire basic data, establish collection analysis transactions, identify valid overdue bills, and generate overdue characteristic information. The profile analysis unit is used to generate the user's original payment profile based on the user's payment behavior and reminder contact behavior within a preset historical time window, and to determine the user's natural payment baseline based on natural payment samples. The collection strategy unit is used to determine the collection priority and reach matching comprehensive amount based on valid overdue bills, overdue characteristic information, user's original payment profile and user's natural payment baseline, generate differentiated collection strategies and form strategy intervention records; The feedback determination unit is used to determine the impact of the collection intervention based on the strategy intervention record, collection reach feedback information and payment feedback information, and to determine the credibility status of the feedback. The closed-loop update unit is used to update the user payment profile, save the effect sample of the collection strategy, or postpone the update based on the feedback credibility status, and form a collection closed loop status.

Images