Customer service optimization method and system based on generative artificial intelligence

Abstract

The invention discloses a customer service optimization method and system based on generative artificial intelligence, and relates to the technical field of artificial intelligence, and the method comprises the steps: reading the retention parcel attribute and retention duration, and setting a reminding short message sending frequency and configuring a sending time zone in combination with the site load and the historical average pickup duration of a user; retrieving a sample data set from a parcel operation database based on the user portrait and a preset parcel taking association factor; training a generative artificial intelligence model by using the sample data, and constructing a pickup probability predictor; and in the short message sending time zone, a predictor is used for optimizing and determining an optimal time node corresponding to the maximum predicted pickup probability, and a reminding short message is sent to a user. Therefore, the technical effects of accurate reminding, improvement of processing efficiency and improvement of user experience are achieved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a customer service optimization method and system based on generative artificial intelligence. Background Art

[0002] In the postal service industry, SMS notifications are a common method for handling stranded packages. Traditionally, SMS notifications are triggered at predetermined, fixed times to inform users of their packages' detention. While this method can provide notifications to a certain extent, it has significant limitations and struggles to meet the increasingly complex needs of package management and personalized user experiences.

[0003] Existing technologies for SMS reminders of stranded packages are unable to make real-time and precise adjustments based on numerous dynamic factors, such as the package's actual stranded status, package attributes, site load, and user behavior. Specifically, regardless of the package's attributes, the site's current load, or the user's past pickup habits, SMS reminders are sent at fixed, uniform intervals. This often results in inaccurate reminder timing. Often, SMS messages may be sent when it's inconvenient for the user to pick up their package, or factors such as the site's processing capacity and package priority may not be fully considered, making effective reminders impossible. This impacts package processing efficiency and reduces the user experience. Summary of the Invention

[0004] The present invention provides a customer service optimization method and system based on generative artificial intelligence to solve the technical problems in the existing technology of insufficient accuracy of reminder timing, affecting processing efficiency and user experience, and achieve the technical effects of accurate reminders, improved processing efficiency and improved user experience.

[0005] In a first aspect, the present invention provides a customer service optimization method based on generative artificial intelligence, wherein the customer service optimization method based on generative artificial intelligence includes:

[0006] Read the package attributes and detention duration of the stranded package, set the real-time sending frequency of reminder SMS based on the site load and the user's historical average pickup time, and configure the SMS sending time zone.

[0007] Using user profiles as comparison constraints, information retrieval is performed in the parcel operation database according to the preset pickup correlation factors to obtain a sample data set.

[0008] The sample data set is used as training data to train a generative artificial intelligence model and build a pickup probability predictor.

[0009] The pickup probability predictor is used to optimize the SMS sending time within the SMS sending time zone to determine the optimal time node, and a reminder SMS is sent to the user at the optimal time node, wherein the optimal time node is the time node corresponding to the maximum predicted pickup probability.

[0010] In a feasible implementation, the package attributes include package volume, item value, and storage environment constraints.

[0011] In a feasible implementation, the setting of the real-time sending frequency of the reminder SMS and the configuration of the SMS sending time zone include:

[0012] The number of packages at the current site is read, and the ratio of the number of packages at the current site to the preset number threshold is set as the site load.

[0013] The urgency of the pickup is assessed based on the package volume, item value, storage environment constraints, detention time, site load, and historical average pickup time of users to determine the urgency of the pickup.

[0014] The preset SMS sending frequency is compensated and rounded according to the urgency of the pickup to obtain the real-time sending frequency, and the SMS sending time zone is configured based on the real-time sending frequency.

[0015] In one feasible implementation, the urgency of pickup is assessed based on the package volume, item value, storage environment constraints, detention time, station load, and historical average pickup time of users, including:

[0016] Using the coefficient of variation method and combined with historical parcel operation data, we weighted the package volume, item value, storage environment constraints, detention time, station load, and the historical average user pickup time. Based on the weighted configuration results, we constructed an urgency evaluation function, in which the pickup urgency is positively correlated with the package volume, item value, storage environment constraints, detention time, station load, and the historical average user pickup time.

[0017] After dimensionless processing of the package volume, item value, storage environment constraints, detention time, site load, and historical average pickup time of users, the urgency of pickup is evaluated and determined using the urgency evaluation function.

[0018] In one feasible implementation, user profiles are used as comparison constraints, and information is retrieved from the parcel operation database according to preset pickup correlation factors to obtain a sample data set, including:

[0019] Obtain a user profile, wherein the user profile includes basic attribute characteristics and receiving and sending behavior characteristics.

[0020] Configure preset pickup correlation factors, wherein the preset pickup correlation factors include at least time nodes, weather conditions and the number of packages to be picked up at the site.

[0021] Using the user profile as a comparison constraint, information retrieval is performed in the parcel operation database according to the preset pickup correlation factors to obtain a sample correlation factor set, and the pickup ratios of different sample correlation factors within the preset time window are counted and set as the sample pickup probability to obtain a sample pickup probability set.

[0022] The sample association factor set and the sample pickup probability set are mapped and combined to obtain the sample data set.

[0023] In one feasible implementation, the sample dataset is used as training data to train a generative artificial intelligence model and construct a pickup probability predictor, including:

[0024] The sample data set is divided into K equal parts, and K parts are selected with replacement to construct a first training set. The sample data set is then iteratively selected K times to obtain K training sets, where K is an integer greater than 5.

[0025] The K training sets are used to train generative artificial intelligence models respectively until the models converge, thereby obtaining K pickup probability prediction branches, which are combined to obtain the pickup probability predictor.

[0026] In a feasible implementation, the pickup probability predictor is used to optimize the SMS sending time within the SMS sending time zone to determine the optimal time node, including:

[0027] The SMS sending time zone is divided according to a preset time window to determine a plurality of candidate sending times.

[0028] Based on the preset pickup correlation factors, correlation information of the multiple candidate sending times within a future preset time window is collected respectively to obtain multiple real-time correlation factors.

[0029] The pickup probability predictor is used to predict the pickup probabilities of the multiple real-time correlation factors respectively, output multiple predicted pickup probabilities, and set the candidate sending time corresponding to the maximum predicted pickup probability as the optimal time node.

[0030] In a feasible implementation, using the pickup probability predictor to predict the pickup probability for each of the multiple real-time correlation factors includes:

[0031] Configure the number P of selected pickup probability prediction branches according to the pickup urgency.

[0032] A first real-time correlation factor is randomly selected from the multiple real-time correlation factors.

[0033] P pickup probability prediction branches are randomly selected from the K pickup probability prediction branches of the pickup probability predictor, and the pickup probability is predicted for the first real-time correlation factor respectively to obtain P predicted pickup probabilities, the first predicted pickup probability is obtained by average calculation, and multiple predicted pickup probabilities are analyzed in sequence.

[0034] In a feasible implementation, the number P of selected pickup probability prediction branches is configured according to the pickup urgency, including:

[0035] The ratio of the pickup urgency to the preset pickup urgency is set as the branch adjustment coefficient.

[0036] The product of the branch adjustment coefficient and the initial number of branches is rounded to an integer to obtain the selected number P, wherein the initial number of branches is 2.

[0037] In a second aspect, the present invention further provides a customer service optimization system based on generative artificial intelligence, wherein the customer service optimization system based on generative artificial intelligence includes:

[0038] The SMS configuration module is used to read the package attributes and detention duration of the stranded package, set the real-time sending frequency of the reminder SMS based on the site load and the user's historical average pickup time, and configure the SMS sending time zone.

[0039] The information retrieval module is used to perform information retrieval in the parcel operation database based on the user profile as a comparison constraint and the preset pickup correlation factors to obtain a sample data set.

[0040] Generate a training module for training a generative artificial intelligence model using the sample data set as training data to build a pickup probability predictor.

[0041] The prediction and sending module is used to use the pickup probability predictor to optimize the SMS sending time within the SMS sending time zone to determine the optimal time node, and send a reminder SMS to the user at the optimal time node, wherein the optimal time node is the time node corresponding to the maximum predicted pickup probability.

[0042] The present invention discloses a customer service optimization method and system based on generative artificial intelligence, comprising: obtaining attribute information of pending stranded packages and their stranded duration, and dynamically calculating the sending frequency of reminder text messages in combination with the operating load of the current site and the historical average pickup time of users, and accordingly defining the time period for sending text messages; based on user portrait data, performing information retrieval in the parcel operation database according to set pickup-related factors, and extracting a sample data set for model training; using the sample data set to train a generative artificial intelligence model, and constructing a pickup probability prediction model for specific users; with the help of the prediction model, performing probability optimization analysis within a preset text message sending time period, and screening out the time point with the highest corresponding pickup probability as the optimal time node for text message sending; sending a reminder text message to the target user at the optimal time node, and improving the actual pickup response rate after the text message is reached. The customer service optimization method and system based on generative artificial intelligence disclosed by the present invention solve the technical problems of insufficient accuracy of reminder timing, which affects processing efficiency and user experience, and achieves the technical effects of accurate reminders, improved processing efficiency, and improved user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0043] Figure 1 Schematic diagram of the customer service optimization method based on generative artificial intelligence of the present invention.

[0044] Figure 2 This is a structural diagram of the customer service optimization system based on generative artificial intelligence of the present invention.

[0045] Description of the accompanying drawings: SMS configuration module 11, information retrieval module 12, generation training module 13, prediction and sending module 14. DETAILED DESCRIPTION

[0046] The above technical solution will be described in detail below in conjunction with the accompanying drawings and specific implementation methods of the specification to better understand the above technical solution. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments of the present invention. It should be understood that the present invention is not limited to the example embodiments used only to explain the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention. In addition, it should be noted that, for the convenience of description, only the parts related to the present invention, rather than all, are shown in the drawings.

[0047] Example 1, as Figure 1 The figure is a flow chart of a customer service optimization method based on generative artificial intelligence according to the present invention, wherein the customer service optimization method based on generative artificial intelligence includes:

[0048] S100: Read the package attributes and detention duration of the detained package, set the real-time sending frequency of the reminder SMS based on the site load and the user's historical average pickup time, and configure the SMS sending time zone.

[0049] Specifically, a stranded package refers to a package that fails to be delivered to the recipient in a timely manner according to normal procedures during the logistics process due to various reasons (such as delivery delays, unknown addresses, recipients failing to pick up the package in time, etc.).

[0050] Specifically, the package attributes and detention duration of the detained package are read. These attributes include package volume, item value, and storage environment constraints. These attributes comprehensively reflect the characteristics and storage requirements of the package. Detention duration refers to the length of time the package remains at the station. Furthermore, the station load (the relationship between the station's current package capacity and processing capacity) and the average historical user pickup time (the average time interval reflected by past user pickup behavior) must also be considered.

[0051] By combining these factors, you can set the real-time SMS frequency (the number of times SMS messages are sent within a certain period) and configure the SMS sending time zone (the time range or time period during which SMS messages are sent). For example, if a station currently has a large number of packages, a heavy load, and a relatively long average pickup time for users, you can set a lower real-time SMS frequency and configure the SMS sending time zone to a period during station operating hours and when user activity is high, such as 8:00 PM to 10:00 PM on weekdays. This ensures that users receive reminders at a convenient time while also avoiding excessive SMS sending during busy stations that could disrupt station staff.

[0052] In some embodiments, the package attributes include package volume, item value, and storage environment constraints.

[0053] Optional package attributes include package volume, item value, and storage environment constraints. Package volume influences storage space and transportation options; item value affects the importance of the package and potential insurance requirements; and storage environment constraints refer to any special requirements for storage conditions, such as moisture and shock resistance.

[0054] In some embodiments, the step of setting the real-time sending frequency of the reminder SMS and configuring the SMS sending time zone includes:

[0055] The number of packages at the current site is read, and the ratio of the current number of packages at the site to a preset number threshold is set as the site load; the urgency of pickup is evaluated based on the package volume, item value, storage environment constraints, detention time, site load, and the user's historical average pickup time to determine the pickup urgency; the preset SMS sending frequency is compensated and rounded according to the pickup urgency to obtain the real-time sending frequency, and the SMS sending time zone is configured based on the real-time sending frequency.

[0056] Specifically, "Pickup Urgency" is a comprehensive, quantitative indicator of the urgency of a package's current pickup, based on multiple parameters such as package attributes (including package volume, item value, and storage environment constraints), package retention time, station load, and historical average user pickup time. "Site Load" refers to the ratio of the current number of packages at a station to a preset threshold, reflecting the real-time status of the station's storage pressure and processing capacity. "SMS Sending Time Zone" refers to the time interval within a day used to allocate SMS reminder delivery times, with its width dynamically determined by the SMS sending frequency.

[0057] Specifically, the current station load is calculated by first reading the number of packages at the station and then comparing this number to the preset package threshold to determine the current station load. For example, if the current number of packages is 300 and the threshold is 500, the station load is 0.6. Simultaneously, the system reads information such as the target package's volume, value, storage environment constraints, duration of detention, and the corresponding user's historical average pickup time. Based on these parameters, a weighted or model-based algorithm is used to assess the urgency of package pickup. For example, larger, more valuable, or packages with special storage requirements have a higher urgency for the same detention time. The higher the station load, the greater the urgency weighting coefficient.

[0058] Furthermore, the preset SMS frequency is compensated and rounded up using the pickup urgency factor to obtain the final real-time SMS frequency. For example, if the preset frequency is once per day and the urgency compensation factor is 1.5, the final frequency is rounded up to twice per day. Finally, the 24-hour day is divided into a corresponding number of equal-width time zones based on the real-time sending frequency. For example, if the sending frequency is twice per day, each 12-hour zone will contain SMS reminders, with SMS reminders sent once in the morning and once in the afternoon.

[0059] Through this process, we achieve dynamic and adaptive adjustments to the stranded package reminder policy. We can optimize the frequency and timeframe of SMS reminders in real time based on the actual attributes of the package, site load, and user behavior, effectively improving the timeliness of user pickups and reducing the risk of long-term package delays. Furthermore, properly configuring SMS delivery time zones prevents users from receiving invalid reminders during inactive periods, improving user experience and site operational efficiency.

[0060] In some implementations, the urgency of pickup is assessed based on the package volume, item value, storage environment constraints, detention time, station load, and historical average pickup time of users, including:

[0061] Using the coefficient of variation method and combined with historical parcel operation data, the package volume, item value, storage environment constraints, detention time, station load and historical average user pickup time are weighted. An urgency evaluation function is constructed based on the weight configuration results, where the pickup urgency is positively correlated with the package volume, item value, storage environment constraints, detention time, station load and historical average user pickup time. After dimensionless processing of the package volume, item value, storage environment constraints, detention time, station load and historical average user pickup time, the urgency evaluation function is used to evaluate and determine the pickup urgency.

[0062] Specifically, the coefficient of variation method, combined with historical parcel operation data, was used to assign weights to six factors: parcel volume, item value, storage constraints, detention time, station load, and historical average user pickup time. Based on historical data, the coefficient of variation for each factor was calculated. Factors with larger coefficients of variation were given greater weights, as this indicates that the factor varies significantly across different situations and has a more critical impact on the urgency of pickup. For example, if the coefficient of variation for parcel volume is 0.3, the item value is 0.5, and the storage constraints are 0.6, then the storage constraints have the highest weight.

[0063] Furthermore, an urgency evaluation function is constructed based on the above weights, and the form of the function can be the sum of the products of each factor and its weight.

[0064] Specifically, each of the six indicators is dimensionlessly processed, for example, using min-max normalization to scale each indicator value to the range [0, 1]. These values are then substituted into the urgency evaluation function to calculate the urgency of the package's pickup. A higher urgency value indicates a more urgent pickup need.

[0065] Through the above process, the urgency of each package's pickup can be accurately quantified. This allows for differentiated processing based on the urgency of each package when subsequently configuring the real-time frequency and time zones for SMS reminders. For packages with high urgency, the frequency can be increased appropriately, and reminders can be prioritized during time periods that are more consistent with user pickup habits (i.e., SMS delivery time zones). This effectively encourages users to pick up their packages promptly, reduces package delays, improves package processing efficiency, and ultimately enhances the user experience.

[0066] S200: Using the user profile as a comparison constraint, information is retrieved from the parcel operation database according to the preset pickup correlation factors to obtain a sample data set.

[0067] Specifically, user profiles are constructed based on basic user attributes (such as age, gender, and occupation) and their delivery and collection behaviors (such as pickup frequency and preferred pickup time). These models comprehensively describe user characteristics and reflect their personalized needs and behavior. During information retrieval, user profiles serve as conditional constraints (i.e., comparison constraints) to ensure that retrieved information matches the user's characteristics, thereby improving retrieval accuracy and relevance.

[0068] Specifically, pre-set pickup correlation factors are a variety of factors related to user pickup behavior, such as time of day, weather conditions, and the number of packages waiting for pickup at a station. These factors may affect the likelihood of a user picking up their package. The parcel operations database stores a large amount of data related to parcel operations, including basic package information, user delivery and collection records, and station operational status. It serves as a fundamental data source for information retrieval and data analysis. The sample dataset is a collection of data retrieved and filtered from the parcel operations database that matches the user profile and pre-set pickup correlation factors and is used for subsequent model training.

[0069] This step allows us to accurately filter data from massive amounts of parcel operation data to identify those that match specific user profiles and pre-set pickup factors, generating a high-quality sample dataset. This provides a rich and accurate data foundation for subsequent training of generative AI models, enabling them to better learn the patterns and characteristics of user pickup behavior.

[0070] In some embodiments, using user profiles as comparison constraints, information retrieval is performed in the parcel operation database according to preset pickup correlation factors to obtain a sample data set, including:

[0071] Obtain a user profile, wherein the user profile includes basic attribute characteristics and collection and delivery behavior characteristics; configure preset pickup association factors, wherein the preset pickup association factors include at least time nodes, weather conditions, and the number of packages to be picked up at the site; use the user profile as a comparison constraint, perform information retrieval in the parcel operation database according to the preset pickup association factors, obtain a sample association factor set, and count the pickup ratios of different sample association factors within a preset time window, set them as sample pickup probabilities, and obtain a sample pickup probability set; map and combine the sample association factor set and the sample pickup probability set to obtain the sample data set.

[0072] Specifically, basic attribute characteristics refer to basic information such as the user's identity background, such as age, gender, occupation, and residential area. Collection and delivery behavior characteristics refer to the user's behavior patterns when collecting and sending packages, such as the frequency of collection, the time period for collection, and the preferred collection station. For example, collection and delivery behavior characteristics include:

[0073] Table 1 Exemplary characteristics of receiving and sending behaviors

[0074] Feature Name Explanation of meaning Average pickup time The average time it takes for users to pick up their packages from the time the packages arrive at the site Pickup time preference Common pickup times for users (e.g., weekday evenings, weekend daytimes, etc.) Pickup frequency The number of packages the user picks up per week or month Package type preference The types of products that users frequently receive (such as fresh food, electronic products, books, etc.) Historical retention rate The proportion of user packages exceeding the set detention threshold Frequently used sites User's preferred delivery location Complaint / Feedback Records Do users frequently complain about package delays, service issues, etc.?

[0075] Specifically, time points refer to time factors related to user pickup behavior, including specific hours, dates, and days of the week, such as weekday evenings or weekend afternoons. Weather conditions refer to the weather conditions in the user's area during the pickup period, such as sunny, rainy, and snowy days. Different weather conditions may affect the user's willingness to go out to pick up their parcels. The number of packages waiting to be picked up at a specific station refers to the number of packages waiting for users to pick up at a specific station. This value may affect the convenience and willingness of users to pick up their parcels. For example, too many packages at a station may cause users to spend more time looking for their parcels.

[0076] Specifically, the sample correlation factor set is a data set related to preset pickup correlation factors retrieved from a database, and each sample correlation factor represents various influencing factors in a specific pickup scenario.

[0077] Specifically, we first acquire a user profile encompassing basic attributes and collection and delivery behavior characteristics to fully understand the user's profile. Next, we configure preset pickup correlation factors, including at least time points, weather conditions, and the number of packages waiting for pickup at the station. Then, using the user profile as a comparison constraint, we perform an information search within the parcel operations database. Using database query techniques, we match the user profile with user records in the database, filtering out eligible data to form a sample correlation factor set. Finally, we calculate the pickup ratio for each sample correlation factor within a preset time window to obtain a sample pickup probability set.

[0078] For example, in the past month, for a user, under the conditions of 8:00 PM to 10:00 PM on weekdays, sunny weather, and fewer than 50 packages waiting to be picked up at the station, the number of successful pickups was 8, and the number of eligible packages encountered was 10. Therefore, the corresponding sample pickup probability is 80%. Finally, the sample correlation factor set and the sample pickup probability set are mapped and combined to obtain the sample dataset. For example, the above sample correlation factor (8:00 PM to 10:00 PM on weekdays, sunny weather, and fewer than 50 packages waiting to be picked up at the station) and the corresponding sample pickup probability of 80% are combined into a data record. Multiple such records constitute the sample dataset.

[0079] Through the above process, based on user portraits and multi-dimensional pickup correlation factors, it is possible to accurately mine and quantify the actual probability of users' pickup behavior in different scenarios, forming a personalized, data-driven behavior analysis sample data set, providing support for subsequent pickup behavior prediction, intelligent reminder strategy formulation, user stratification operations and other applications, helping to improve package flow efficiency and user service experience.

[0080] S300: Using the sample data set as training data, training a generative artificial intelligence model and building a pickup probability predictor.

[0081] In some embodiments, the sample dataset is used as training data to train a generative artificial intelligence model and construct a pickup probability predictor, including:

[0082] The sample data set is divided into K equal parts, and K parts are selected with replacement to construct a first training set. The first part is then iteratively selected K times to obtain K training sets, where K is an integer greater than 5. The generative artificial intelligence model is trained separately using the K training sets until the model converges, thereby obtaining K pickup probability prediction branches, which are combined to obtain the pickup probability predictor.

[0083] Specifically, a "generative AI model" is a machine learning model that can generate or predict outputs based on input features. It is preferably a deep neural network, a generative adversarial network (GAN), or other generative models suitable for behavioral probabilistic modeling. The "Pickup Probability Predictor" is a generative AI model trained specifically for the pickup probability prediction task. It infers the input user profile and pickup-related factors to determine the probability of a user picking up a package in a specific scenario.

[0084] Specifically, the sample dataset obtained above is first divided into K equal parts (K is an integer greater than 5, preferably 10). Using sampling with replacement, a number of samples from the K parts are randomly selected each time to construct a first training set. This sampling process is repeated K times to obtain K overlapping training sets. A generative AI model is then trained based on each training set. Exemplarily, an adaptive loss function such as cross-entropy loss or mean squared error is used during training until the model converges, resulting in K independent pickup probability prediction branches. These K pickup probability prediction branches are then integrated, for example, using voting, averaging, or weighted ensemble methods, to construct a final pickup probability predictor. This predictor can output a corresponding pickup probability prediction value for any input user profile and pickup-related factor.

[0085] Through the above process, sampling with replacement and multi-model integration effectively improve the generalization and robustness of the pickup probability predictor, reducing the risk of overfitting that can occur with a single model. The resulting pickup probability predictor can accurately predict pickup probabilities based on different user profiles and scenario-related factors, providing highly reliable decision support for applications such as intelligent reminders and operational optimization.

[0086] S400: Utilize the pickup probability predictor to optimize the SMS sending time within the SMS sending time zone to determine the optimal time node, and send a reminder SMS to the user at the optimal time node, wherein the optimal time node is the time node corresponding to the maximum predicted pickup probability.

[0087] Specifically, a text message sending time zone refers to a pre-defined time range or time period within which text messages can be sent. This range is determined based on multiple factors, such as user behavior and site operating hours, and is used to limit the possible time range for text message delivery. Within this time zone, a pickup probability predictor can be used to find the optimal time to send text messages, ensuring that the text messages sent are most likely to encourage users to pick up their items.

[0088] In some embodiments, using the pickup probability predictor, optimizing the SMS sending time within the SMS sending time zone to determine the optimal time node includes:

[0089] The SMS sending time zone is divided according to a preset time window to determine a plurality of candidate sending times; based on a preset pickup correlation factor, the correlation information of the plurality of candidate sending times in a future preset time window is collected respectively to obtain a plurality of real-time correlation factors; the pickup probability predictor is used to predict the pickup probability of the plurality of real-time correlation factors respectively, and a plurality of predicted pickup probabilities are outputted, and the candidate sending time corresponding to the maximum predicted pickup probability is set as the optimal time node.

[0090] Specifically, the preset time window refers to a pre-set SMS sending time granularity interval (e.g., every 10 minutes, every 30 minutes) based on business needs or user behavior characteristics, which is used to divide the SMS sending time zone into multiple candidate time periods. The preset pickup correlation factor refers to the relevant variables that affect the user's pickup behavior, including but not limited to the user's historical pickup time, geographic location, weather, holidays, courier type, user preferences, etc., which are used to assist in predicting the user's pickup probability within a certain time period. The real-time correlation factor is based on the multi-dimensional dynamic data related to the user's pickup behavior collected in real time within the preset time window in the future based on each candidate sending time node.

[0091] Specifically, the SMS sending time zone (e.g., 8:00 AM - 8:00 PM) is first divided into several candidate sending time nodes (e.g., 8:00 AM, 8:15 AM, 8:30 AM, ..., 8:00 PM) according to preset time windows (e.g., every 15 minutes). Each node is a candidate SMS sending time. Then, for each candidate sending time node, real-time data for that time node within a future preset time window is collected based on preset pickup correlation factors. For example, historical user pickup probabilities during that time period, weather conditions, remaining locker capacity, and holiday information are collected to generate multiple real-time correlation factors.

[0092] Furthermore, a pickup probability predictor is used to predict the pickup probabilities of these multiple real-time correlation factors, output multiple predicted pickup probabilities, and set the candidate sending time corresponding to the maximum predicted pickup probability as the optimal time node.

[0093] For example, the pickup probability predictor predicts that the pickup probability is 0.6 when the SMS is sent at 8:00, the pickup probability is 0.7 when it is sent at 8:30, and the pickup probability is 0.65 when it is sent at 9:00. Then 8:30 can be set as the optimal time node and a reminder SMS can be sent to the user at this time.

[0094] Through the above process, the optimal time to send SMS reminders can be accurately determined within the SMS sending time zone. By fully considering various factors that influence user pickup behavior, such as weather and package volume at the station, as well as their actual conditions at different times, and combining this with a pickup probability predictor, the effectiveness of SMS reminders can be maximized, ensuring that users receive reminders at the most likely pickup time, thereby improving pickup efficiency. This also significantly optimizes the package handling process, increases package processing speed, and enhances the user experience. Users receive reminders at the appropriate time, reducing package delays and increasing service satisfaction.

[0095] In some implementations, using the pickup probability predictor to predict the pickup probability for each of the multiple real-time correlation factors includes:

[0096] The number P of selected pickup probability prediction branches is configured according to the urgency of pickup; a first real-time correlation factor is randomly selected from the multiple real-time correlation factors; P pickup probability prediction branches are randomly selected from the K pickup probability prediction branches of the pickup probability predictor, and the pickup probability is predicted for the first real-time correlation factor respectively to obtain P predicted pickup probabilities, the first predicted pickup probability is obtained by average calculation, and multiple predicted pickup probabilities are analyzed in sequence.

[0097] Specifically, based on the preset mapping relationship or corresponding rules, according to the current user's urgency of picking up the item, the number of pickup probability prediction branches P (P≤K) required for this prediction is automatically configured. That is, if the urgency is high, P takes a larger value, and if the urgency is low, P takes a smaller value.

[0098] Furthermore, one of the multiple real-time correlation factors to be analyzed is randomly selected as the first real-time correlation factor. Then, P branches are randomly selected from the K prediction branches of the pickup probability predictor. The first real-time correlation factor is input into each of these P branches to obtain P pickup probability prediction results. Finally, the average of these P prediction results is taken as the final predicted pickup probability for the first real-time correlation factor.

[0099] According to the above steps, the branch selection, prediction and mean calculation processes can be repeated for the remaining real-time correlation factors in turn, and finally multiple predicted pickup probabilities can be obtained.

[0100] The above-mentioned method significantly improves the robustness and generalization of pickup probability predictions through multi-branch integrated prediction and mean fusion, reducing the error caused by fluctuations in single-branch predictions. Dynamically adjusting the number of branches based on pickup urgency helps balance prediction efficiency and accuracy, achieving intelligent and flexible customer service optimization.

[0101] In some implementations, configuring the number P of selected pickup probability prediction branches according to the pickup urgency includes:

[0102] The ratio of the pickup urgency to the preset pickup urgency is set as the branch adjustment coefficient; the product of the branch adjustment coefficient and the initial number of branches is rounded to an integer to obtain the selection number P, where the initial number of branches is 2.

[0103] Specifically, the preset pickup urgency refers to a pre-set standard urgency threshold, serving as a reference for urgency grading and branch adjustment. The branch adjustment coefficient is the ratio of the current pickup urgency to the preset pickup urgency, a parameter used to dynamically adjust the number of predicted branches. The initial number of branches is the default number of branches predicted for the minimum pickup probability, typically 2.

[0104] Specifically, first, calculate the ratio of the current pickup urgency to the preset pickup urgency to obtain the branch adjustment coefficient. For example, if the current urgency is 4 and the preset urgency is 2, the branch adjustment coefficient is 4 / 2 = 2. Then, multiply the branch adjustment coefficient by the initial number of branches (e.g., 2) and round the result to the integer to obtain the number of branches P to be selected for this prediction. For example, if the branch adjustment coefficient is 2 and the initial number of branches is 2, then P = 2 × 2 = 4.

[0105] Optionally, an upper limit is set for P according to the actual situation of the application environment to prevent excessive number of branches from wasting computing resources.

[0106] The above process dynamically adjusts the number of branches by the ratio of the pickup urgency to the preset urgency, achieving an adaptive balance between prediction accuracy and efficiency in different urgency scenarios.

[0107] In summary, the customer service optimization method based on generative artificial intelligence provided by the present invention has the following technical effects:

[0108] By obtaining the attribute information of the stranded packages to be processed and their detention duration, and combining the current site's operating load with the user's historical average pickup time, the frequency of sending reminder text messages is dynamically calculated, and the time period for sending text messages is defined accordingly; based on user portrait data, information retrieval is performed in the parcel operation database according to the set pickup-related factors, and a sample data set for model training is extracted; the generative artificial intelligence model is trained using the sample data set to construct a pickup probability prediction model for specific users; with the help of this prediction model, probability optimization analysis is performed within the preset text message sending time period, and the time point with the highest corresponding pickup probability is selected as the optimal time node for text message sending; reminder text messages are sent to target users at the optimal time node to improve the actual pickup response rate after the text message is received, thereby achieving the technical effects of accurate reminders, improved processing efficiency, and improved user experience.

[0109] Example 2, as Figure 2 This is a schematic diagram of the structure of the customer service optimization system based on generative artificial intelligence of the present invention. For example, Figure 1 The flow chart of the customer service optimization method based on generative artificial intelligence in the present invention can be shown as follows: Figure 2 The structure shown is implemented.

[0110] Based on the same concept as the customer service optimization method based on generative artificial intelligence in the above embodiment, the present invention also provides a customer service optimization system based on generative artificial intelligence, including:

[0111] The SMS configuration module 11 is used to read the package attributes and detention duration of the detained package, set the real-time sending frequency of the reminder SMS based on the site load and the user's historical average pickup time, and configure the SMS sending time zone.

[0112] The information retrieval module 12 is used to perform information retrieval in the parcel operation database based on the user profile as a comparison constraint and the preset pickup correlation factors to obtain a sample data set.

[0113] A training module 13 is generated, which is used to train a generative artificial intelligence model using the sample data set as training data to construct a pickup probability predictor.

[0114] The prediction and sending module 14 is used to use the pickup probability predictor to optimize the SMS sending time within the SMS sending time zone to determine the optimal time node, and send a reminder SMS to the user at the optimal time node, wherein the optimal time node is the time node corresponding to the maximum predicted pickup probability.

[0115] In some embodiments, the SMS configuration module 11 includes:

[0116] The site load calculation unit is used to read the number of packages at the current site and set the ratio of the current number of packages at the site to a preset number threshold as the site load.

[0117] The pickup urgency assessment unit is used to assess the pickup urgency based on the package volume, item value, storage environment constraints, detention time, site load and historical average pickup time of users to determine the pickup urgency.

[0118] The SMS sending frequency configuration unit is used to compensate and round off the preset SMS sending frequency according to the urgency of the pickup to obtain the real-time sending frequency, and configure the SMS sending time zone based on the real-time sending frequency.

[0119] In some implementations, the steps executed by the pickup urgency assessment unit in the SMS configuration module 11 include:

[0120] Using the coefficient of variation method and combined with historical parcel operation data, we weighted the package volume, item value, storage environment constraints, detention time, station load, and the historical average user pickup time. Based on the weighted configuration results, we constructed an urgency evaluation function, in which the pickup urgency is positively correlated with the package volume, item value, storage environment constraints, detention time, station load, and the historical average user pickup time.

[0121] After dimensionless processing of the package volume, item value, storage environment constraints, detention time, site load, and historical average pickup time of users, the urgency of pickup is evaluated and determined using the urgency evaluation function.

[0122] In some embodiments, the information retrieval module 12 includes:

[0123] The user portrait acquisition unit is used to acquire a user portrait, wherein the user portrait includes basic attribute characteristics and receiving and sending behavior characteristics.

[0124] The preset pickup correlation factor configuration unit is used to configure the preset pickup correlation factor, wherein the preset pickup correlation factor at least includes the time node, weather conditions and the number of packages to be picked up at the site.

[0125] The sample correlation factor retrieval and pickup probability statistics unit is used to use the user profile as a comparison constraint, perform information retrieval in the parcel operation database according to the preset pickup correlation factor, obtain a sample correlation factor set, and count the pickup ratios of different sample correlation factors within a preset time window, set them as sample pickup probabilities, and obtain a sample pickup probability set.

[0126] The sample data set generating unit is used to map and combine the sample association factor set and the sample pickup probability set to obtain the sample data set.

[0127] In some embodiments, generating the training module 13 includes:

[0128] The sample data set equal division and training set construction unit is used to divide the sample data set into K equal parts, select K times with replacement to construct the first training set, and iteratively select K times to obtain K training sets, where K is an integer greater than 5.

[0129] The generative artificial intelligence model training and predictor construction unit is used to use the K training sets to train the generative artificial intelligence model separately until the model converges, obtain K pickup probability prediction branches, and combine them to obtain the pickup probability predictor.

[0130] In some embodiments, the prediction and sending module 14 includes:

[0131] The unit for determining the time to be sent is used to divide the time zone for sending the short message according to a preset time window and determine a plurality of time to be sent.

[0132] The real-time correlation factor collection unit is used to collect correlation information of the multiple candidate sending times within a future preset time window based on the preset pickup correlation factors to obtain multiple real-time correlation factors.

[0133] The optimal time node determination unit is used to use the pickup probability predictor to predict the pickup probability of the multiple real-time correlation factors respectively, output multiple predicted pickup probabilities, and set the candidate sending time corresponding to the maximum predicted pickup probability as the optimal time node.

[0134] In some implementations, the optimal time node determination unit in the prediction and sending module 14 performs the following steps:

[0135] Configure the number P of selected pickup probability prediction branches according to the pickup urgency.

[0136] A first real-time correlation factor is randomly selected from the multiple real-time correlation factors.

[0137] P pickup probability prediction branches are randomly selected from the K pickup probability prediction branches of the pickup probability predictor, and the pickup probability is predicted for the first real-time correlation factor respectively to obtain P predicted pickup probabilities, the first predicted pickup probability is obtained by average calculation, and multiple predicted pickup probabilities are analyzed in sequence.

[0138] In some implementations, the optimal time node determination unit in the prediction and sending module 14 may further execute the following steps:

[0139] The ratio of the pickup urgency to the preset pickup urgency is set as the branch adjustment coefficient.

[0140] The product of the branch adjustment coefficient and the initial number of branches is rounded to an integer to obtain the selected number P, wherein the initial number of branches is 2.

[0141] It should be understood that the embodiments mentioned in this specification focus on their differences from other embodiments. The specific embodiments in the aforementioned embodiment one are also applicable to the customer service optimization system based on generative artificial intelligence described in embodiment two. For the sake of brevity of the specification, they will not be further elaborated here.

[0142] It should be understood that the embodiments disclosed in the present invention and the above description can enable those skilled in the art to use the present invention to implement the present invention. At the same time, the present invention is not limited to the embodiments mentioned above. It should be understood that those skilled in the art can still modify the technical solutions described in the above embodiments or replace some of the technical features therein with equivalents; and such modifications or replacements do not deviate from the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention and are all included in the scope of protection of the present invention.

Claims

1. A customer service optimization method based on generative artificial intelligence, characterized in that: include: Read the parcel attributes and detention duration of the stranded parcel, set the real-time sending frequency of reminder SMS based on the site load and the user's historical average pickup time, and configure the SMS sending time zone; Using user profiles as comparison constraints, we search the parcel operation database for information based on preset pickup correlation factors to obtain a sample data set. Using the sample data set as training data, training a generative artificial intelligence model to construct a pickup probability predictor; The pickup probability predictor is used to optimize the SMS sending time within the SMS sending time zone to determine the optimal time node, and a reminder SMS is sent to the user at the optimal time node, wherein the optimal time node is the time node corresponding to the maximum predicted pickup probability.

2. The customer service optimization method based on generative artificial intelligence according to claim 1 is characterized in that: The package attributes include package volume, item value, and storage environment constraints.

3. The customer service optimization method based on generative artificial intelligence according to claim 2 is characterized in that: The setting of the real-time sending frequency of reminder SMS and the configuration of the SMS sending time zone include: Read the number of packages at the current site, and set the ratio of the current number of packages at the site to a preset number threshold as the site load; Perform a pickup urgency assessment based on the package volume, item value, storage environment constraints, detention time, station load, and historical average pickup time of users to determine the pickup urgency; The preset SMS sending frequency is compensated and rounded according to the urgency of the pickup to obtain the real-time sending frequency, and the SMS sending time zone is configured based on the real-time sending frequency.

4. The customer service optimization method based on generative artificial intelligence according to claim 3 is characterized in that: The urgency of the pickup is assessed based on the package size, item value, storage environment constraints, detention time, station load, and historical average pickup time of users, including: Using the coefficient of variation method and combined with historical parcel operation data, we weighted the package volume, item value, storage environment constraints, detention time, station load, and the average historical user pickup time. Based on the weighted configuration results, we constructed an urgency assessment function. The urgency of pickup is positively correlated with package volume, item value, storage environment constraints, detention time, station load, and the average historical user pickup time. After dimensionless processing of the package volume, item value, storage environment constraints, detention time, site load, and historical average pickup time of users, the urgency of pickup is evaluated and determined using the urgency evaluation function.

5. The customer service optimization method based on generative artificial intelligence according to claim 1 is characterized in that: Using user profiles as comparison constraints, we search the parcel operation database for information based on pre-set pickup correlation factors to obtain a sample data set, including: Obtaining a user profile, wherein the user profile includes basic attribute characteristics and receiving and sending behavior characteristics; Configure preset pickup correlation factors, where the preset pickup correlation factors include at least time nodes, weather conditions, and the number of packages waiting to be picked up at the station; Using the user profile as a comparison constraint, information is retrieved from the parcel operation database according to the preset pickup correlation factors to obtain a sample correlation factor set, and the pickup ratios of different sample correlation factors within a preset time window are counted, which are set as sample pickup probabilities to obtain a sample pickup probability set; The sample association factor set and the sample pickup probability set are mapped and combined to obtain the sample data set.

6. The customer service optimization method based on generative artificial intelligence according to claim 1, characterized in that: Using the sample dataset as training data, a generative artificial intelligence model is trained to construct a pickup probability predictor, including: Divide the sample data set into K equal parts, select K parts with replacement to construct a first training set, and iterate and select K parts K times to obtain K training sets, where K is an integer greater than 5; The K training sets are used to train generative artificial intelligence models respectively until the models converge, thereby obtaining K pickup probability prediction branches, which are combined to obtain the pickup probability predictor.

7. The customer service optimization method based on generative artificial intelligence according to claim 6, characterized in that: Utilizing the pickup probability predictor, optimizing the SMS sending time within the SMS sending time zone to determine the optimal time node, including: Dividing the SMS sending time zone according to a preset time window to determine multiple candidate sending times; Based on the preset pickup correlation factors, correlation information of the multiple selected sending times within the future preset time window is collected to obtain multiple real-time correlation factors; The pickup probability predictor is used to predict the pickup probabilities of the multiple real-time correlation factors respectively, output multiple predicted pickup probabilities, and set the candidate sending time corresponding to the maximum predicted pickup probability as the optimal time node.

8. The customer service optimization method based on generative artificial intelligence according to claim 7 is characterized in that: Using the pickup probability predictor, respectively predicting the pickup probability for the multiple real-time correlation factors includes: Configure the number of selected pickup probability prediction branches P according to the pickup urgency; randomly selecting a first real-time correlation factor from the plurality of real-time correlation factors; P pickup probability prediction branches are randomly selected from the K pickup probability prediction branches of the pickup probability predictor, and the pickup probability is predicted for the first real-time correlation factor respectively to obtain P predicted pickup probabilities, the first predicted pickup probability is obtained by average calculation, and multiple predicted pickup probabilities are analyzed in sequence.

9. The customer service optimization method based on generative artificial intelligence according to claim 8, characterized in that: Configure the number of pickup probability prediction branches P based on the pickup urgency, including: The ratio of the pickup urgency to the preset pickup urgency is set as the branch adjustment coefficient; The product of the branch adjustment coefficient and the initial number of branches is rounded to an integer to obtain the selected number P, wherein the initial number of branches is 2.

10. A customer service optimization system based on generative artificial intelligence, characterized by: A customer service optimization method based on generative artificial intelligence for implementing any one of claims 1 to 9, comprising: The SMS configuration module is used to read the package attributes and detention duration of the stranded package, set the real-time sending frequency of the reminder SMS based on the site load and the historical average pickup time of users, and configure the SMS sending time zone; The information retrieval module is used to retrieve information from the parcel operation database based on the user profile as a comparison constraint and the preset pickup correlation factors to obtain a sample data set; Generate a training module for training a generative artificial intelligence model using the sample data set as training data to construct a pickup probability predictor; The prediction and sending module is used to use the pickup probability predictor to optimize the SMS sending time within the SMS sending time zone to determine the optimal time node, and send a reminder SMS to the user at the optimal time node, wherein the optimal time node is the time node corresponding to the maximum predicted pickup probability.

Images