Information processing method and device, storage medium and program product

Abstract

The embodiment of the invention provides an information processing method and device, a storage medium and a program product. In the embodiment of the invention, after the intelligent customer service enters the trusteeship state, the question information and the subsequent question information received within the validity period of the distributed lock are stored in the message queue as a current batch of question information needing delay processing, the delay processing can give real communication feeling to merchants, and the service experience of the merchants is improved. Furthermore, under the condition that timing of the timing task of the message queue is finished, an AI question and answer service model based on artificial intelligence is called to perform unified reply processing on the question information currently needing delay processing in the message queue, so that target reply information corresponding to the obtained question information is sent to a client of a target merchant; and the target merchant can check the target reply information corresponding to each piece of question information, so that manual customer service is replaced, the labor is saved, and the question processing efficiency is improved.

Description

Technical Field

[0001] This application relates to the field of computer processing technologies, and in particular, to an information processing method, device, storage medium, and program product. Background Art

[0002] Currently, some service platforms provide services for promoting merchant posts. Merchants have a large number of problems to be solved by dedicated customer service every day. These tasks rely heavily on manual customer service to handle, occupying a large amount of time of manual customer service, and the response efficiency is relatively low.

[0003] How to quickly and effectively solve the problems of merchants, appease the emotions of merchants, promote the healthy development of the platform, and bring a better service experience to merchants is an urgent problem to be solved by the service platform currently. Summary of the Invention

[0004] Multiple aspects of this application provide an information processing method, device, storage medium, and program product, which can save labor, improve the efficiency of problem-solving, and at the same time provide a certain delay processing service, give merchants a real conversation feeling, and improve the service experience of merchants.

[0005] An embodiment of this application provides an information processing method. When receiving the question information of a target merchant corresponding to an intelligent customer service in a managed state, attempt to obtain a distributed lock corresponding to the target merchant, and the distributed lock has a validity period; if the distributed lock corresponding to the target merchant is successfully obtained, create a message queue corresponding to the distributed lock, and the distributed lock and the message queue cooperate to collect the current batch of question information that needs to be delayed for processing; store the question information and subsequent question information received within the validity period of the distributed lock into the message queue to obtain the current batch of question information that needs to be delayed for processing; the message queue corresponds to a timing task, and the timing task is used to time the delay processing time, and the validity period of the distributed lock is greater than or equal to the delay processing time; in the case where the timing task ends timing, call an AI question-answering service model based on artificial intelligence to perform unified reply processing on the current batch of question information stored in the message queue to obtain target reply information corresponding to each question information; send the target reply information corresponding to each question information to the client of the target merchant for the target merchant to view the target reply information corresponding to each question information.

[0006] An embodiment of this application also provides an electronic device, including: a memory and a processor; the memory is used to store a computer program; the processor is coupled to the memory and is used to execute the computer program to implement the steps in the above method.

[0007] An embodiment of the present application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to implement the steps in the above method.

[0008] An embodiment of the present application also provides a computer program product, which includes a computer program / instructions that, when executed by a processor, enable the processor to implement the steps in the above method.

[0009] In an embodiment of the present application, a hosting service for intelligent customer service and a delayed processing service for question information are provided. After the intelligent customer service enters the hosting state, if question information of the target merchant corresponding to the intelligent customer service is received, an attempt is made to obtain a distributed lock corresponding to the target merchant and create a message queue corresponding to the distributed lock after successful acquisition, so as to store the question information and subsequent question information received within the validity period of the distributed lock into the message queue as the current batch of question information that needs to be delayed for processing. Delayed processing can give merchants a real communication experience and improve the service experience of merchants. Further, when the timing task of the message queue ends, an AI question-and-answer service model based on artificial intelligence is called to uniformly reply to the question information that needs to be delayed for processing in the message queue, so as to send the target reply information corresponding to each question information to the client of the target merchant for the target merchant to view the target reply information corresponding to each question information, replacing the manual customer service, saving labor, and improving the processing efficiency of problems. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The drawings described herein are used to provide a further understanding of the present application and form a part of the present application. The illustrative embodiments of the present application and their descriptions are used to explain the present application and do not constitute an improper limitation of the present application. In the drawings:

[0011] Figure 1 It is a schematic flowchart of an information processing method provided by an exemplary embodiment of the present application;

[0012] Figure 2a It is a schematic flowchart of another information processing method provided by another exemplary embodiment of the present application;

[0013] Figure 2b It is a schematic flowchart of another information processing method provided by yet another exemplary embodiment of the present application;

[0014] Figure 2c It is a schematic flowchart of another information processing method provided by still another exemplary embodiment of the present application;

[0015] Figure 3 It is a schematic structural diagram of an electronic device provided by an exemplary embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0016] To make the objectives, technical solutions, and advantages of this application clearer, the technical solutions of this application will be clearly and completely described below in conjunction with specific embodiments of this application and the corresponding drawings. Obviously, the described embodiments are only a part of the embodiments of this application, rather than all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of this application.

[0017] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Moreover, the collection, use, and processing of relevant data need to comply with relevant laws, regulations, and standards of relevant countries and regions, and corresponding operation entrances are provided for users to choose to authorize or refuse.

[0018] In response to the above technical problems, in the embodiments of this application, a hosting service and a delay processing service for intelligent customer service are provided. After the intelligent customer service enters the hosting state, if question information of the target merchant corresponding to the intelligent customer service is received, an attempt is made to obtain the distributed lock corresponding to the target merchant, and after the acquisition is successful, a message queue corresponding to the distributed lock is created to store the question information and subsequent question information received within the validity period of the distributed lock into the message queue as the current batch of question information that needs to be delayed for processing. Delayed processing can give merchants a real communication experience and improve the service experience of merchants. Further, when the timing task of the message queue ends, an AI question and answer service model based on artificial intelligence is called to uniformly reply to the question information that needs to be delayed for processing in the message queue, so as to send the target reply information corresponding to each question information to the client of the target merchant for the target merchant to view the target reply information corresponding to each question information, replacing the manual customer service, saving labor, and improving the processing efficiency of problems.

[0019] The following will describe in detail a solution provided by the embodiments of this application in conjunction with the drawings.

[0020] Figure 1 It is a schematic flowchart of an information processing method provided for an exemplary embodiment of this application. Figure 2a It is a schematic flowchart of another information processing method provided for an exemplary embodiment of this application. As Figure 1 and as shown in 2a, it includes:

[0021] 101. When question information of the target merchant corresponding to the intelligent customer service in the hosting state is received, an attempt is made to obtain the distributed lock corresponding to the target merchant, and the distributed lock has a validity period;

[0022] 102. If the distributed lock corresponding to the target merchant is successfully obtained, create a message queue corresponding to the distributed lock. The distributed lock and the message queue are used in cooperation to collect the current batch of question information that needs to be processed later.

[0023] 103. Store the question information and the subsequent question information received within the validity period of the distributed lock into the message queue to obtain the current batch of question information that needs to be processed later. The message queue corresponds to a timing task, and the timing task is used to time the delay processing time, and the validity period of the distributed lock is greater than or equal to the delay processing time.

[0024] 104. When the timing task ends, call the AI Q&A service model based on artificial intelligence to uniformly reply to the current batch of question information stored in the message queue to obtain the target reply information corresponding to each question information.

[0025] 105. Send the target reply information corresponding to each question information to the client of the target merchant for the target merchant to view the target reply information corresponding to each question information.

[0026] Generally, some applications provide customer service. Customer service is a service (question-and-answer service) provided by the application for merchants or users to communicate with the customer service. Among them, the user refers to a user who searches for the services or products he needs from the target application. A merchant refers to a merchant who cooperates with the application program and promotes the services or goods it can provide in the target application. The cooperation method between the merchant and the application can be, for example, that the merchant purchases the membership of the application and can enjoy the membership rights and interests after becoming a member of the application. The membership rights and interests can be, for example, the peace-of-mind investment right, the daily booth right, the precise service right, etc. Among them, the peace-of-mind investment right is a promotion service, mainly used to help merchants improve the exposure rate and traffic of posts, and it can help merchants quickly obtain more high-quality traffic through functions such as intelligent product selection and intelligent keyword selection; the daily booth right is a post promotion product purchased by the day, which can meet the needs of merchants' posts for 24-hour online promotion throughout the day; the precise right is a digital marketing plan, usually used to help merchants achieve precise reach to target customers, and it can provide more efficient promotion effects for merchants through big data analysis and intelligent algorithms. The following embodiments of the present application focus on the question-and-answer service between the merchant and the target application.

[0027] In this embodiment, the type of the target application is not limited. The target application can be any APP running on a terminal device. For example, it can be a housekeeping APP, a house viewing APP, a transportation APP, a live broadcast APP, and so on. Alternatively, the target application can also be a web page or a mini program running on an APP. In addition, the type of the terminal device is not limited in this embodiment. The terminal device can be a smart handheld device, such as a smart phone, a tablet computer, and so on. Alternatively, the terminal device can also be a desktop device, such as a laptop computer, a desktop computer, and so on; or the terminal device can also be a smart wearable device, such as a smart watch, a smart bracelet, and so on; or the terminal device can also be various smart home appliances with a display screen, such as a smart TV, a smart large screen, or a smart robot, and so on.

[0028] Generally, the target application replies to the question information of the merchant through a human customer service. To improve the reply efficiency and save labor, a smart customer service is introduced, and the smart customer service can quickly give a reply message according to the user's question information.

[0029] Further optionally, to enhance the real communication experience between the merchant and the smart customer service, a hosting node can be introduced for the customer service. The hosting node provides a hosting service for hosting the smart customer service. When a merchant corresponding to a smart customer service in the hosting state asks a question, the question and the questions asked within a certain period of time will be stored with a delay to obtain the current batch of questions that need to be processed with a delay. When the delay duration arrives, unified reply processing will be performed on the current batch of questions that need to be processed with a delay. Optionally, the hosting node can be implemented as middleware. Middleware is an independent system software or service program located between the operating system and the application program, providing general services for the application program, but the implementation form of the hosting node is not limited to this.

[0030] In an alternative embodiment, the hosting service corresponds to a hosting service page, and the hosting service page includes: a hosting QR code, and the hosting QR code supports scanning operations. When a smart customer service needs to be hosted, the human customer service staff corresponding to the smart customer service can host the corresponding smart customer service through the hosting QR code on the hosting service page. Further optionally, the hosting service page can be displayed in response to a display request for the hosting service page initiated by the human customer service staff. The hosting service page includes a hosting QR code; in response to a successful scanning operation of the hosting QR code by the human customer service staff, the smart customer service corresponding to the human customer service staff is determined and assigned to the hosting customer service queue. The smart customer service in the hosting customer service queue pauses providing customer service for the corresponding merchant.

[0031] Optionally, in response to the successful scanning operation of the managed QR code by the human customer service staff, determine the intelligent customer service corresponding to the human customer service staff and allocate the intelligent customer service to the managed customer service queue, including: both the human customer service staff and the intelligent customer service corresponding to the human customer service staff have unique identification information, then in response to the successful scanning operation of the managed QR code by the human customer service staff, the identification information of the human customer service staff can be obtained; according to the identification information of the human customer service staff, determine the identification information of the intelligent customer service corresponding to the human customer service staff; determine the intelligent customer service according to the identification information of the intelligent customer service and allocate the intelligent customer service to the managed customer service queue.

[0032] Further optionally, a distributed lock management node is also provided for the customer service. The distributed lock management node provides a distributed lock management service for storing and managing the distributed locks corresponding to each merchant. Among them, the distributed lock has a validity period, and the validity periods of different distributed locks can be the same or different. The distributed lock is used to manage the message queue corresponding to the distributed lock, and store the question information of the merchant corresponding to the distributed lock within the validity period of the distributed lock into the message queue, so as to obtain the current batch of question information that needs to be delayed for processing, that is, to realize the delayed processing of the current batch of question information. Optionally, the distributed lock management node can be implemented as middleware. Middleware is an independent system software or service program located between the operating system and the application program, providing general services for the application program, but the implementation form of the distributed lock management node is not limited to this.

[0033] In the embodiment of the present application, when receiving the question information of the target merchant corresponding to the intelligent customer service in the managed state, the managed node attempts to obtain the distributed lock corresponding to the target merchant from the distributed lock management node.

[0034] In an optional embodiment, both the target merchant and each distributed lock have unique identification information, and one merchant can correspond to a dedicated distributed lock. Based on this, the managed node attempts to obtain the distributed lock corresponding to the target merchant from the distributed lock management node, including: sending a distributed lock acquisition request to the distributed lock management node, where the acquisition request contains the identification information of the target merchant; according to the identification information of the target merchant, determine the identification information of the distributed lock corresponding to the target merchant; determine the distributed lock according to the identification information of the distributed lock and send the distributed lock to the managed node.

[0035] Further optionally, each distributed lock stored in the distributed lock management node corresponds to usage status information, and the usage status information includes: usage status and idle status. After successfully acquiring the distributed lock corresponding to the target merchant, the distributed lock management node may mark the distributed lock as "usage status". When the distributed lock is in the "usage status", the distributed lock cannot be acquired again based on the subsequent question information of the target merchant.

[0036] In the embodiment of the present application, if the distributed lock corresponding to the target merchant is successfully acquired, a message queue corresponding to the distributed lock is created. The distributed lock and the message queue cooperate to collect the question information that needs to be delayed for processing in the current batch. Optionally, the message queue may be implemented as a Redis message queue. The Redis message queue is implemented based on the list data structure of Redis. The list is a linear data structure that supports operations from the head (left) and the tail (right). Through these operations, the behavior of a first-in-first-out (FIFO) queue can be simulated. Among them, the first-in-first-out (FIFO) queue means that the queue is a data structure with first-in-first-out, that is, the element that enters the queue earliest will be processed first when subsequent processing is performed. In this embodiment, the high-performance read and write capabilities of Redis are mainly utilized to implement the temporary storage and sequential processing of messages, that is, to temporarily store the question information sent by the merchant to the intelligent customer service for batch storage management within the validity period of the distributed lock.

[0037] In the embodiment of the present application, the message queue corresponds to a timing task, and the timing task is used to time the delay processing time. Optionally, the timing task may have a timing start time and a timing duration. The timing start time of the timing task may be the time when the message queue corresponding to the distributed lock is successfully created. The timing duration refers to the length of the timing time of the timing task. When the timing duration of the timing task meets the duration set by the timing task, the timing task ends. The length of the timing duration in this embodiment is not limited and may be set according to specific delay processing requirements. The timing duration may be a few seconds, 1 minute or several minutes, etc. Optionally, the timing duration of the timing task can be controlled by the hosting node. After the message queue is successfully created, the hosting node sends a delay message to the timing task, and the delay message includes the timing duration.

[0038] In the embodiment of the present application, after the message queue is successfully created, the question information that attempts to acquire the distributed lock corresponding to the target merchant and the subsequent question information of the target merchant received within the validity period of the distributed lock are sequentially stored in the message queue corresponding to the distributed lock to obtain the question information that needs to be delayed for processing in the current batch.

[0039] In an optional embodiment, after successfully obtaining the distributed lock corresponding to the target merchant based on the received question information of the intelligent customer service in the managed state, during the timing of this round of timing tasks, if the question information of the target merchant is received again, the managed node will still attempt to obtain the distributed lock corresponding to the target merchant from the distributed lock management node. When the distributed lock management node queries that the distributed lock is in the "in-use state", it will prompt the managed node that the distributed lock is in the "in-use state" and guide the managed node to store the question information in the message queue corresponding to the distributed lock until the current round of timing tasks corresponding to the message queue ends. That is to say, whenever the question information of the target merchant corresponding to the intelligent customer service in the managed state is received, an attempt will be made to obtain the distributed lock corresponding to the target merchant. If the distributed lock is in the "in-use state", the acquisition will fail. If the distributed lock is in the "idle state", the distributed lock corresponding to the target merchant will generally be successfully obtained.

[0040] In the embodiment of the present application, when the timing of the timing task ends, an AI question-and-answer service model based on artificial intelligence can be called to uniformly reply to the current batch of question information that needs to be delayed and stored in the message queue to obtain the target reply information corresponding to each question information. Among them, the question information can be text, image or video. The AI question-and-answer service model based on artificial intelligence can be trained based on a large amount of historical communication information between merchants and customer service. During the communication based on the AI question-and-answer service model, the AI question-and-answer service model can also be fine-tuned in real time according to the communication information generated during the communication process to continuously improve the accuracy of the reply information of the AI question-and-answer service model. In addition, the AI question-and-answer service model has also learned a lot of knowledge about prompt words, so that during the communication process, it can reply under the guidance of the input prompt words, thereby further improving the accuracy of the AI service question-and-answer model's understanding of the question information and the accuracy of the reply information given.

[0041] In an optional embodiment, when calling the AI question-and-answer service model based on artificial intelligence to uniformly reply to the current batch of question information that needs to be delayed and stored in the message queue, the reply process can be carried out in sequence according to the storage order of the current batch of question information that needs to be delayed and stored in the message queue. Thereby, it is possible to avoid the unrealistic communication experience brought to the merchant by simultaneously giving the reply information of the current batch of question information that needs to be delayed.

[0042] The technical solutions provided by the above embodiments of the present application offer a hosting service and a delayed processing service for intelligent customer service. After the intelligent customer service enters the hosting state, if a question message from the target merchant corresponding to the intelligent customer service is received, an attempt is made to obtain the distributed lock corresponding to the target merchant, and after successful acquisition, a message queue corresponding to the distributed lock is created to store the question message and subsequent question messages received within the validity period of the distributed lock into the message queue as the current batch of question messages that need to be delayed for processing. Delayed processing can give merchants a real communication experience and improve the service experience of merchants. Further, when the timing task of the message queue ends, an AI question and answer service model based on artificial intelligence is called to uniformly reply to the question messages that need to be delayed for processing in the message queue, so as to send the target reply messages corresponding to each question message to the client of the target merchant for the target merchant to view the target reply messages corresponding to each question message, replacing the manual customer service, saving labor, and improving the problem processing efficiency.

[0043] Further optionally, when the timing task ends, that is, when the timing time arrives, the distributed lock corresponding to the target merchant can be released to facilitate the collection of the next batch of question messages that need to be delayed for processing.

[0044] More specifically, when the timing task ends, the hosting node can stop the cooperation between the distributed lock and the message queue. The hosting node releases the distributed lock to the distributed lock management node. Correspondingly, the distributed lock management node modifies the current state of the distributed lock to the "idle state" to facilitate obtaining the distributed lock again from the distributed lock management node when a question message from the target merchant corresponding to the intelligent customer service in the hosting state is received again. Among them, the delayed processing time can be determined by the timing duration of the timing task, that is, the timing duration of the timing task is equivalent to the delayed processing time. In addition, the validity period of the distributed lock is greater than the delayed processing time to ensure that the distributed lock is valid during the delayed processing time.

[0045] In an optional embodiment, an AI Q&A service model based on artificial intelligence is called to uniformly process and reply to a current batch of question information stored in the message queue that needs to be processed with a delay, so as to obtain target reply information corresponding to each piece of question information, including: inputting each piece of question information stored in the message queue and a first model prompt word into the AI Q&A service model, where the first model prompt word is used to prompt the AI Q&A service model to provide intent recognition service and Q&A service; under the prompt of the first model prompt word, for each piece of question information, perform an intent recognition operation to obtain the target intent information of each piece of question information, and the target intent information represents the purpose of the merchant's question; and based on the target intent information of each piece of question information, combined with the historical data of the target merchant related to the target intent information of each piece of question information, generate the target reply information corresponding to each piece of question information. The historical data refers to various data generated during the cooperation process between the merchant and the target application. The historical data can be, for example, data related to various rights and interests (historical posting times, posted times of posts, historical consultation times, used times of various rights and interests, remaining membership amount). In this embodiment, the target reply information is not given only according to the target intent information corresponding to each piece of question information, but the target reply information corresponding to each piece of question information is obtained according to the historical data of the target merchant related to the target intent information of each piece of question information, which can improve the accuracy of the target reply information.

[0046] For example. Taking the merchant's enjoyment of the daily booth right as an example, the daily booth can meet the need for the merchant's post to be promoted online 24 hours a day. If the merchant does not receive any consultation from users after posting a post on the first day of enjoying the daily booth, the merchant can send a question information to the customer service. The question information can be, for example, "Why is there no consultation for the posted post?" After inputting this question into the AI Q&A service model, after the AI service model recognizes the target intent information corresponding to this question, it can directly give a reply information according to the target intent information. The reply information can be, for example, "The promotion process takes time. Please be patient. The longer the promotion time, the more likely there will be more consultations!" Further, after the AI Q&A service model recognizes the target intent information corresponding to this question, it can view the historical data related to the merchant's daily booth right. The historical data can be, for example: the post publishing duration of the daily booth is 2 hours and the current consultation number is 0. Then the reply information given by the AI Q&A service model according to the target intent information and the historical data can be, for example, "The promotion time of your daily booth post is 2 hours and there is indeed no consultation at present. However, the promotion process takes time. Please be patient. The longer the promotion time, the more likely there will be more consultations!" Thus, it can be seen that the AI Q&A service model gives more specific and accurate target reply information corresponding to the question information by combining historical data, which can improve the service experience of the merchant.

[0047] In practical applications, the first model prompt can be, for example, "Please act as an expert in intent recognition, identify the target intent information corresponding to each question information, and when identifying the target intent information corresponding to each question information, first identify the target intent information type corresponding to each question information from multiple pre-set intent types, and after identifying the target intent information type, determine the target intent information from multiple intent information under the pre-set target intent information type", but not limited to this.

[0048] Further optionally, different intent information may correspond to different intent types. Then, for each question information, perform an intent recognition operation to obtain the target intent information of each question information, including: performing an intent classification operation based on the keywords in each question information to determine the target intent information type adapted to each question information from multiple known intent types; for each question information, perform an intent recognition operation under the target intent information type adapted to the question information to identify the target intent information adapted to the question information under the target intent information type. Performing an intent hierarchical recognition operation when identifying the target intent information of each question information, first identifying the target intent information type corresponding to each question information, and then identifying the target intent information from the target intent information type, can improve the accuracy of identifying the target intent information.

[0049] Further optionally, before performing an intent classification operation based on the keywords in each question information to determine the target intent information type adapted to each question information from multiple known intent types, it also includes: determining the keywords of each question information according to TF-IDF (Term Frequency-Inverse Document Frequency); or determining the keywords of each question information based on a rule-based method; or determining the keywords of each question information based on a machine learning method; or determining the keywords of each question information based on a deep learning method. Among them, TF-IDF is a method for measuring the importance of vocabulary, which combines term frequency (TF) and inverse document frequency (IDF). TF (term frequency) refers to the frequency of a word appearing in historical Q&A information, reflecting the importance of the word in historical Q&A information. IDF (inverse document frequency): measures the general importance of a word and is achieved by calculating the rarity of the word in the entire historical Q&A information. The TF-IDF formula is: TF-IDF(t, d) = TF(t, d) × IDF(t), where t is the word and d is the document. The rule-based method includes but is not limited to: stop word filtering method, part-of-speech filtering method, pattern matching method, etc.

[0050] Among them, to determine the keywords of each question information according to TF-IDF, it includes: for each question information, perform word segmentation operation to obtain multiple word segments included in each question information; calculate the TF value of each word segment and calculate the IDF value of each word segment; based on the TF value and IDF value of each word segment, calculate the TF-IDF value of each word segment, and select the word with the highest TF-IDF value as the keyword.

[0051] Among them, when determining the keywords of each question information by the rule-based method, the stop word filtering method is to remove the common and meaningless words in the text (such as "of", "is", "and", etc.), and only retain the words that may be meaningful. Specifically, it can be implemented as: maintaining a stop word list and filtering out these words after word segmentation; the part-of-speech filtering method is to extract keywords according to the part of speech (such as noun, verb, adjective), and these parts of speech usually can express the core content of the text more. Specifically, it can be implemented as: using a part-of-speech tagging tool (such as Stanford NLP, HanLP, etc.) to extract words of specific parts of speech; the pattern matching method is to extract keywords according to the predefined pattern (such as "question word + noun"). Specifically, it is implemented as: defining rules, such as "why + noun", "how + verb", etc.

[0052] Among them, when determining the keywords of each question information by the machine learning-based method, it includes supervised learning methods and unsupervised learning methods. Among them, for the supervised learning method, a model can be trained using the labeled keyword data, and the model learns how to extract keywords from the text. Specifically, it can be implemented as: preparing labeled data (text and its keywords); extracting text features (such as TF-IDF vectors); training the model (such as SVM, Naive Bayes, deep learning model); using the model to predict keywords. For the unsupervised learning method, keywords can be automatically extracted through the structure and semantic information of the text without labeled data. Specifically, it can be implemented as: using the graph-based ranking (TextRank) algorithm, that is, similar to PageRank, constructing a graph through the co-occurrence relationship between words and calculating the importance of words; or, using the topic LDA (Latent Dirichlet Allocation) algorithm, that is, extracting topic words as keywords through the three-layer structure of document-topic-word.

[0053] Among them, when using deep learning-based methods to determine the keywords of each question information, architectures such as WordEmbedding + neural network and Transformer architecture (such as BERT) can be used. Among them, using the Word Embedding + neural network architecture and the Transformer architecture is to convert words into semantic vectors through pre-trained word embedding algorithms (such as Word2Vec, GloVe, BERT), and then extract keywords through neural network algorithms (such as CNN, RNN, Transformer). Specifically, it can be implemented as follows: obtaining the embedding vector of the word using a pre-trained model (such as BERT); inputting it into a neural network model (such as Bi-LSTM, Transformer); using the attention mechanism (Attention) to highlight important words; and outputting keywords. Using the Transformer architecture (such as BERT) is to generate dynamic embeddings of words through the BERT model using context information, which can better capture the semantic information of words. Specifically, it can be implemented as follows: encoding the text using BERT; extracting the output of the hidden layer and calculating the importance of each word; and selecting keywords according to the importance scores.

[0054] Further optionally, multiple intent type data tables can be pre-maintained. Different services correspond to different intent type data tables, and each intent type data table contains multiple intent types. Based on this, after obtaining the keywords in each question information, perform intent classification operations according to the keywords in each question information to determine the target intent information type adapted to each question information from multiple known intent types, including: obtaining the target intent information type data table corresponding to the customer service according to the type of customer service; and obtaining the target intent information type adapted to the keywords in each question information from the target intent information type data table.

[0055] More specifically, obtaining the target intent information type adapted to the keywords in each question information from the target intent information type data table. One optional implementation includes: extracting the semantic feature information of the keywords in each question information, and extracting the semantic feature information of each intent type included in the target intent information type data table; calculating the semantic feature distance between the semantic feature information of the keywords in each question information and the semantic feature information of each intent type included in the target intent information type data table, and taking the intent type with the smallest semantic feature distance as the target intent information type corresponding to the current question information. Alternatively, obtaining the target intent information type adapted to the keywords in each question information from the target intent information type data table. Another optional implementation includes: performing word segmentation on each intent type included in the target intent information type data table to obtain multiple word segments included in each intent type; calculating the semantic feature distance between the keywords in each question information and each word segment in each intent type; taking the intent type corresponding to the smallest semantic feature distance as the target intent information type corresponding to the question information. It can be understood that the closer the semantic feature distance between the keywords in each question information and the intent type, the greater the probability that the intent type is the target intent information type of the question information, and vice versa.

[0056] In practical applications, the intent types can be, for example: promotion questions, phone questions, cost questions, etc. The multiple intent information included in the promotion questions can be, for example: poor promotion effect, promotion operation problems, promotion cost problems, coupon usage problems, etc. The multiple intent information included in the phone questions can be, for example: few phone calls problem, no phone calls problem, phone signal problems, etc. The multiple intent information included in the cost questions can be, for example: promotion cost problems, phone cost problems, etc.

[0057] Further optionally, the multiple intent type data tables maintained in advance also include multiple intent information under each intent type. After obtaining the target intent information type adapted to each question information, for each question information, under the target intent information type adapted to the question information, perform an intent recognition operation to identify the target intent information adapted to the question information under the target intent information type, including: obtaining multiple intent information under the target intent information type from the intent type data table according to the target intent information type; obtaining the intent information adapted to the keywords in each question information from the multiple intent information as the target intent information. For the specific implementation of this embodiment, reference can be made to the relevant description of performing the intent type recognition operation, which will not be elaborated here. It can be understood that the closer the semantic feature distance between the keywords in each question information and the intent information, the greater the probability that the intent information is the target intent information of the question information, and vice versa.

[0058] Further optionally, different intent information has corresponding data types, and different intents correspond to different data types. For example, the "few calls" intent corresponds to relevant data of the "few calls" type, and the "phone signal" intent corresponds to relevant data of the "phone information" type. Additionally, for the same question information, there may be multiple response information. Then, one of the response information needs to be selected as the target response information. For this purpose, the database integrates multiple response information involved in historical communication information and / or multiple response information pre-generated by the model based on historical data. Different response information corresponds to different priorities, so that the AI Q&A service model can determine the target response information corresponding to each question information according to the priorities of each response information. Based on this, based on the target intent information of each question information, combined with the historical data of the target merchant related to the target intent information of each question information, the target response information corresponding to each question information is generated, including: for each question information, determine the target data type related to the target intent information according to the target intent information of the question information; according to the target data type, obtain the historical data under the target data type from the first database; based on the target intent information, obtain multiple response information adapted to the target intent information from the second database as multiple candidate response information; determine the target response information of this question information according to the response priorities and historical data of each of the multiple candidate response information. The first database and the second database can be the same database or different databases. This embodiment focuses on the case where the first database and the second database are different databases. More specifically, the first database can be a relational database or a non-relational database. Examples of relational databases include MySQL, PostgreSQL, Oracle Database, etc., and examples of non-relational databases include Redis, MongoDB, Apache Cassandra, etc.

[0059] Optionally, determining the target response information of this question information according to the response priorities and historical data of each of the multiple candidate response information includes: sequentially determining whether the historical data meets the data conditions included in each candidate response information according to the response priorities of each of the multiple candidate response information; using the candidate response information corresponding to the data condition that matches the historical data and is closest to the historical data as the target response information.

[0060] For example. Taking the merchant's question "Why is there no one consulting on the posts of the daily exhibition booths that have been published?" as an example, the multiple candidate reply messages matched by this question message may include, but are not limited to: Reply message 1: "The promotion process takes time. Please wait patiently. The longer the promotion time, the more likely the consultation volume will be!"; Reply message 2: "The promotion duration of your daily exhibition booth post is currently within 2 hours. There is indeed no one consulting at present, but the promotion process takes time. Please wait patiently. The longer the promotion time, the more likely the consultation volume will be!"; Reply message 3: "The promotion duration of your daily exhibition booth post is currently within 3 hours. There is indeed no one consulting at present, but the promotion process takes time. Please wait patiently. The longer the promotion time, the more likely the consultation volume will be!". Among them, the promoted duration in reply message 3 includes the promoted duration in reply message 2, and the specific promoted duration is not included in reply message 1. Then the accuracy of reply message 2 is higher than that of reply message 3, and the accuracy of reply message 3 is higher than that of reply message 1. Then the priority of reply message 2 is higher than that of reply message 3, and the priority of reply message 3 is higher than that of reply message 1. Then, when determining the target reply message of this question message only based on the order of priorities, the target reply message of this question message should be reply message 2. However, if the "promoted duration is 2.5 hours" is recorded in the historical data, then reply message 2 with the highest priority among the candidate reply messages does not meet the historical data. Determining the target reply message as reply message 2 only based on the priorities of the candidate reply messages does not conform to the promoted duration loaded in the historical data, and thus needs to be excluded.

[0061] Further optionally, the complexity corresponding to each intent information can be pre-maintained. Different intent information has different complexities, and the complexity represents the difficulty level of solving the question messages related to the intent information. The AI question and answer service model can solve the question messages whose complexity is adapted to its capabilities. For some question messages with a relatively high difficulty level, the AI question and answer service model may not be able to solve them and can only be solved by artificial customer service staff. Based on this, before generating the target reply message corresponding to each question message based on the target intent information of each question message and combining the historical data related to the target merchant and the target intent information of each question message, it further includes: identifying the complexity of the target intent information based on the pre-configured complexity level of the intent information; if the AI question and answer service model is adapted to the complexity of the target intent information, then using the AI question and answer service model to obtain the historical data and multiple candidate reply messages related to the target intent information of the target merchant based on the target intent information. Thus, it is possible to avoid providing the question messages that the AI question service model cannot solve to the AI question service model, which affects the reply efficiency.

[0062] Further optionally, the AI Q&A service model can also provide corresponding marketing plans for merchants based on the historical question information of the merchants. Specifically, the second model prompt can be input into the AI Q&A service model, and the second model prompt is used to prompt the AI Q&A service model to provide a timed marketing push service; under the prompt of the second model prompt, at least one target intention information corresponding to the question information of each merchant within a preset time period is obtained, and a marketing plan adapted to the at least one target intention information is obtained; the marketing plan adapted to the at least one target intention information is sent to the corresponding client so that the corresponding merchant can view the marketing plan.

[0063] Optionally, under the prompt of the second model prompt, at least one target intention information corresponding to the question information of each merchant within a preset time period is obtained, and a marketing plan adapted to the at least one target intention information is obtained, including: using the intention information ranked before a preset ranking corresponding to the question information of each merchant within the preset time period as at least one target intention information; according to the at least one target intention information, marketing plans adapted to each target intention information are respectively obtained from the marketing plan database. Further, the marketing plans can be pushed in sequence according to the magnitude order of the probability values of the target intention information corresponding to each marketing plan; or, for each marketing plan adapted to the obtained target intention information, a plan merging process is performed to obtain a target marketing plan, so as to push the target marketing plan to the client. Among them, the preset time period can be, for example, within several hours, the same day, this week, this month, etc. The preset ranking can be, for example, the first place, the second place, the third place, etc.

[0064] In practical applications, the second model prompt can be, for example, "Please act as an expert in recommending marketing plans. According to the first target intention information and the second target intention information ranked in the top N of the question information of the target merchant within this week, select the first marketing plan adapted to the first target intention information and the second marketing plan adapted to the second target intention information from the marketing database", but not limited to this. Taking the question information "Why is there no consultation for the posted post" as an example, the determined intention type adapted to this question information can be "promotion problem", the target intention information can be "poor promotion effect", and the corresponding marketing plan can be, for example, "Based on your question information this week and the historical data related to the question information, it is recommended that you add 1 more promotional post and purchase 1-day post active push service to actively push your post to users and see the effect!"

[0065] To facilitate the understanding of the above solution, the following Figure 2b expounds on this solution.

[0066] The artificial customer service personnel log in to the hosting node and host their corresponding intelligent customer service (customer service) through the hosting node (hosting service).

[0067] Further, after the target merchant corresponding to the intelligent customer service in the managed state sends a question message, when the managed node receives the question message (receives the message), the AI Q&A service model (algorithm) will determine whether it can identify the corresponding target intent information based on the question information (whether it can be identified), and if the target intent information is identified, historical data will be obtained from the business (main site) related to the target intent information. During the process of obtaining historical data, the historical data is verified to check whether the obtained historical data is data related to the target intent information.

[0068] In the case where it is determined that the corresponding target intent information cannot be identified based on the question information, the AI Q&A service model acts as the intelligent customer service and informs the target merchant that manual customer service personnel need to handle it, so that the target merchant can ask questions to the manual customer service personnel.

[0069] Further, if the verification result shows that the obtained historical data is data related to the target intent information corresponding to the question information, the AI Q&A service model obtains the target reply information (obtains the answer content) adapted to the target intent information. After obtaining the target reply information, the AI Q&A service model acts as the intelligent customer service and sends the target reply information to the target merchant through the managed ID of the intelligent service corresponding to the target merchant or the ID of the target merchant corresponding to the intelligent customer service. The conversation between the target merchant and the intelligent customer service role can be carried out through the chat page provided by the target application or through the enterprise instant messaging application corresponding to the intelligent customer service (such as WeCom). It should be noted that Figure 2b the step of "whether it can be identified" is actually included in the step of "obtaining the answer content", but in the figure, in order to distinguish each step, "whether it can be identified" is marked outside "obtaining the answer content".

[0070] If the verification result shows that the obtained historical data is not relevant to the target intent information corresponding to the question information, the customer service process ends.

[0071] During this process, if the manual customer service personnel take over the question information of the target merchant, the intelligent customer service will stop being managed for a period of time, such as 30 minutes, to leave enough time for the manual customer service personnel to communicate with the target merchant.

[0072] To facilitate the understanding of the above solution, the following combines exemplary question information and Figure 2c elaborates on this solution.

[0073] Exemplary question information of the target merchant corresponding to the intelligent customer service in the managed state can be, for example, "The information consulted by the user when calling does not match the services or goods provided in the post". Input this question information into the AI Q&A service model. The AI Q&A service model acts as the intelligent customer service and conducts multiple rounds of Q&A with the target merchant to collect business opportunity information. The business opportunity information can be, for example, the ID of the relevant post (business opportunity ID), the detailed information of the post, the detailed information of the target merchant, the mobile phone numbers (phone numbers) of the user and the target merchant, and the call date between the user and the target merchant.

[0074] Further, based on the obtained business opportunity information, the AI Q&A service model obtains the deduction situation of the mobile phone number of the target merchant from the historical information of the relevant service. In the process of the AI Q&A service model obtaining the deduction situation of the mobile phone number of the target merchant from the historical information of the relevant service based on the obtained business opportunity information, it will judge whether the target merchant has conducted a transaction with the user who communicated with it by phone (whether the business opportunity channel has received an order). If the judgment result is yes, it will continue to judge whether the target merchant has continued to follow up this transaction.

[0075] If it is judged that the target merchant has not conducted a transaction with the user who communicated with it by phone, it will continue to judge whether it is a valid phone number for An Xin Tou. If the judgment result is no, it will be determined that this deduction is invalid.

[0076] Further, after the AI Q&A service model obtains the deduction situation, if it is judged that the deduction is an invalid deduction based on the above process, it will send a soothing message to the target merchant to soothe the target merchant's emotions. If necessary, the deducted fee this time can be refunded to the target merchant. After the AI Q&A service model obtains the deduction situation, if it is judged that the deduction is a valid deduction based on the above process, it will prompt the customer service through a system message that this deduction is valid (a system message will be directly sent to the customer service to create an online record for a valid deduction).

[0077] Figure 3 The following is a schematic structural diagram of an electronic device provided by an exemplary embodiment of the present application. As Figure 3 shown, it includes: a memory 30a and a processor 30b; the memory 30a is used to store a computer program; the processor 30b is coupled to the memory 30a and is used to execute the computer program to implement the following steps:

[0078] When receiving the question information of the target merchant corresponding to the intelligent customer service in the escrow state, attempt to obtain the distributed lock corresponding to the target merchant. The distributed lock has a validity period. If the distributed lock corresponding to the target merchant is successfully obtained, create a message queue corresponding to the distributed lock. The distributed lock and the message queue are used in combination to collect the current batch of question information that needs to be processed later. Store the question information and the subsequent question information received within the validity period of the distributed lock into the message queue to obtain the current batch of question information that needs to be processed later. The message queue corresponds to a timing task, and the timing task is used to time the delay processing time, and the validity period of the distributed lock is greater than or equal to the delay processing time. When the timing task ends, call the AI question and answer service model based on artificial intelligence to perform unified reply processing on the current batch of question information stored in the message queue to obtain the target reply information corresponding to each question information. Send the target reply information corresponding to each question information to the client of the target merchant for the target merchant to view the target reply information corresponding to each question information.

[0079] Further, the processor is further configured to: when the timing task ends, release the distributed lock corresponding to the target merchant to facilitate the collection of the next batch of question information that needs to be processed later.

[0080] In some embodiments, when the processor calls the AI question and answer service model based on artificial intelligence to perform unified reply processing on the current batch of question information stored in the message queue to obtain the target reply information corresponding to each question information, it is specifically configured to: input each question information stored in the message queue and the first model prompt word into the AI question and answer service model. The first model prompt word is used to prompt the AI question and answer service model to provide intent recognition service and question and answer service. Under the prompt of the first model prompt word, for each question information, perform an intent recognition operation to obtain the target intent information of each question information. And based on the target intent information of each question information, combine the historical data related to the target intent information of the target merchant and each question information to generate the target reply information corresponding to each question information.

[0081] Optionally, when the processor performs an intent recognition operation on each question information to obtain the target intent information of each question information, it is specifically configured to: perform an intent classification operation according to the keywords in each question information to determine the target intent information type adapted to each question information from multiple known intent types. For each question information, perform an intent recognition operation under the target intent information type adapted to the question information to identify the target intent information adapted to the question information under the target intent information type.

[0082] Optionally, when generating the target response information corresponding to each question message based on the target intent information of each question message and combining the historical data of the target merchant related to the target intent information of each question message, the processor is specifically configured to: for each question message, determine the target data type related to the target intent information according to the target intent information of the question message; obtain the historical data of the target data type from the first database according to the target data type; obtain multiple response information adapted to the target intent information from the second database based on the target intent information as multiple candidate response information; and determine the target response information of the question message according to the response priorities and historical data of the multiple candidate response information.

[0083] Optionally, when determining the target response information of the question message according to the response priorities and historical data of the multiple candidate response information, the processor is specifically configured to: sequentially determine whether the historical data meets the data conditions included in each candidate response information according to the response priorities of the multiple candidate response information; and use the candidate response information corresponding to the data condition that matches the historical data and is closest to the historical data as the target response information.

[0084] Further optionally, before generating the target response information corresponding to each question message based on the target intent information of each question message and combining the historical data of the target merchant related to the target intent information of each question message, the processor is further configured to: identify the complexity of the target intent information based on the pre-configured complexity level of the intent information; if the AI question and answer service model is adapted to the complexity of the target intent information, use the AI question and answer service model to obtain the historical data and multiple candidate response information related to the target intent information of the target merchant based on the target intent information.

[0085] Further optionally, the processor is further configured to: display a hosting service page, where the hosting service page includes: a hosting QR code; in response to a successful scan operation of the hosting QR code by the customer service staff, determine the intelligent customer service corresponding to the customer service staff and assign the intelligent customer service to the hosting customer service queue.

[0086] Further optionally, the processor is further configured to: input the second model prompt word into the AI question and answer service model, where the second model prompt word is used to prompt the AI question and answer service model to provide a timed marketing push service; under the prompt of the second model prompt word, obtain a marketing plan adapted to at least one target intent information according to the at least one target intent information corresponding to the question messages of each merchant within a preset time period; and send the marketing plan adapted to at least one target intent information to the corresponding client so that the corresponding merchant can view the marketing plan.

[0087] The detailed implementation manners and beneficial effects of each module in the embodiments of the present application have been described in detail in the foregoing embodiments, and will not be elaborated herein.

[0088] Further, as Figure 3 shown, the electronic device further includes: other components such as a communication component 30c, a display 30d, a power supply component 30e, an audio component 30f, etc. Figure 3 Only some components are schematically shown in Figure 3 and it does not mean that the electronic device only includes Figure 3 the components shown. Additionally, Figure 3 the components within the dashed box in Figure 3 are optional components, rather than essential components, and can be determined according to the product form of the electronic device. The electronic device in this embodiment can be implemented as a terminal device such as a desktop computer, a laptop computer, a smart phone or an IOT device, or can also be a server device such as a conventional server, a cloud server or a server array. If the electronic device in this embodiment is implemented as a terminal device such as a desktop computer, a laptop computer, a smart phone, etc., it can include Figure 3 the components within the dashed box in Figure 3 ; if the electronic device in this embodiment is implemented as a server device such as a conventional server, a cloud server or a server array, it may not include Figure 3 the components within the dashed box in

[0089] The above-mentioned memory can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as Static Random-Access Memory (SRAM), Electrically Erasable Programmable Read Only Memory (EEPROM), Erasable Programmable Read Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic memory, flash memory, a magnetic disk or an optical disc.

[0090] The above communication component is configured to facilitate communication between the device where the communication component is located and other devices in a wired or wireless manner. The device where the communication component is located can access a communication standard-based wireless network, such as a WiFi, 2G, 3G, 4G / LTE, 5G, or other mobile communication networks, or a combination thereof. In an exemplary embodiment, the communication component receives a broadcast signal or broadcast-related information from an external broadcast management system via a broadcast channel. In an exemplary embodiment, the communication component further includes a Near Field Communication (NFC) module to facilitate short-range communication. For example, the NFC module can be implemented based on Radio Frequency Identification (RFID) technology, Infrared Data Association (IrDA) technology, Ultra Wide Band (UWB) technology, Bluetooth (BT) technology, and other technologies.

[0091] The above display includes a screen, and the screen can include a Liquid Crystal Display (LCD) and a Touch Panel (TP). If the screen includes a touch panel, the screen can be implemented as a touch screen to receive input signals from a user. The touch panel includes one or more touch sensors to sense touches, swipes, and gestures on the touch panel. The touch sensors can sense not only the boundaries of touch or swipe actions but also detect the duration and pressure associated with the touch or swipe operation.

[0092] The above power component provides power for various components of the device where the power component is located. The power component can include a power management system, one or more power sources, and other components associated with generating, managing, and distributing power for the device where the power component is located.

[0093] The above audio component can be configured to output and / or input audio signals. For example, the audio component includes a microphone (MIC), and when the device where the audio component is located is in an operating mode, such as a call mode, a recording mode, and a voice recognition mode, the microphone is configured to receive external audio signals. The received audio signals can be further stored in a memory or transmitted via the communication component. In some embodiments, the audio component further includes a speaker for outputting audio signals.

[0094] Correspondingly, an exemplary embodiment of the present application further provides a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to implement the steps in the above method.

[0095] Correspondingly, an embodiment of the present application further provides a computer program product, which includes computer programs / instructions. When the computer programs / instructions are executed by a processor, the processor is enabled to implement the steps in the above method.

[0096] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can take the form of a computer program product implemented on one or more computer-readable storage media (including but not limited to disk memory, compact disc read-only memory (CD-ROM), optical memory, etc.) that contain computer-usable program code.

[0097] The present application is described with reference to the flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowchart and / or block diagram, as well as the combination of flows and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, such that the instructions executed by the processor of the computer or other programmable data processing devices generate a device for implementing the specified functions in Figure 1 one or more of the processes Figure 1 or multiple processes and / or blocks

[0098] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, such that the instructions stored in the computer-readable memory generate a manufactured article including an instruction device that implements the specified functions in Figure 1 one or more of the processes Figure 1 or multiple processes and / or blocks

[0099] These computer program instructions can also be loaded onto a computer or other programmable data processing device, such that a series of operation steps are executed on the computer or other programmable device to generate a computer-implemented process. Thus, the instructions executed on the computer or other programmable device provide steps for implementing the specified functions in Figure 1 one or more of the processes Figure 1 or multiple processes and / or blocks

[0100] In a typical configuration, a computing device includes one or more processors (Central Processing Unit, CPU), an input / output interface, a network interface, and memory.

[0101] The memory may include non-permanent memory in the form of computer-readable media, random access memory (Random Access Memory, RAM) and / or non-volatile memory such as read-only memory (ROM) or flash memory (flash RAM). Memory is an example of computer-readable media.

[0102] Computer-readable media includes permanent and non-permanent, removable and non-removable media and can store information by any method or technology. The information can be computer-readable instructions, data structures, program modules, or other data. Examples of computer storage media include, but are not limited to, phase-change random access memory (Phase-change Random Access Memory, PRAM), static random access memory (SRAM), dynamic random access memory (Dynamic Random Access Memory, DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other memory technologies, compact disc read-only memory (CD-ROM), digital versatile disc (Digital Video Disc, DVD) or other optical storage, magnetic cassette tapes, magnetic tape disk storage or other magnetic storage devices, or any other non-transmission media that can be used to store information accessible by a computing device. As defined herein, computer-readable media does not include transitory computer-readable media such as modulated data signals and carrier waves.

[0103] It should also be noted that the term "comprising", "including" or any other variant thereof is intended to cover non-exclusive inclusion, such that a process, method, commodity or device comprising a series of elements not only includes those elements but also includes other elements not expressly listed, or also includes elements inherent to such process, method, commodity or device. Without further limitation, an element defined by the statement "comprising one..." does not exclude the presence of additional identical elements in the process, method, commodity or device comprising the element.

[0104] The above are only embodiments of the present application and are not used to limit the present application. For those skilled in the art, various changes and modifications can be made to the present application. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included within the scope of the claims of the present application.

Claims

1. An information processing method, characterized in that: include: When receiving the question information of the target merchant corresponding to the intelligent customer service in the managed state, try to obtain the distributed lock corresponding to the target merchant, and the distributed lock has a validity period; If the distributed lock corresponding to the target merchant is successfully obtained, a message queue corresponding to the distributed lock is created, and the distributed lock and the message queue are used together to collect the current batch of question information that needs to be delayed; The question information and subsequent question information received within the validity period of the distributed lock are stored in the message queue to obtain a current batch of question information that needs to be delayed; the message queue corresponds to a timing task, the timing task is used to time the delayed processing time, and the validity period of the distributed lock is greater than or equal to the delayed processing time; When the timing task ends, the AI ​​question-answering service model based on artificial intelligence is called to uniformly reply to the current batch of question information that needs to be delayed and stored in the message queue, so as to obtain the target reply information corresponding to each question information; The target answer information corresponding to each question information is sent to the client of the target merchant, so that the target merchant can view the target answer information corresponding to each question information.

2. The method according to claim 1, characterized in that: Also includes: When the timing task ends, the distributed lock corresponding to the target merchant is released to facilitate the collection of the next batch of question information that needs to be delayed.

3. The method according to claim 1, characterized in that Calling the AI ​​question-answering service model based on artificial intelligence to uniformly reply to the current batch of question information that needs to be delayed and stored in the message queue, so as to obtain the target reply information corresponding to each question information, including: Input each question information and the first model prompt word stored in the message queue into the AI ​​question and answer service model, wherein the first model prompt word is used to prompt the AI ​​question and answer service model to provide intent recognition service and question and answer service; Under the prompt of the first model prompt word, for each question information, perform an intent recognition operation to obtain the target intent information of each question information; and based on the target intent information of each question information, combined with the historical data related to the target merchant and the target intent information of each question information, generate the target reply information corresponding to each question information.

4. The method according to claim 3, characterized in that For each question information, perform intent recognition operations to obtain the target intent information of each question information, including: According to the keywords in each question information, an intent classification operation is performed to determine a target intent information type adapted to each question information from a plurality of known intent types; For each of the question information, an intent recognition operation is performed under the target intent information type adapted to the question information to identify the target intent information adapted to the question information under the target intent information type.

5. The method according to claim 3, characterized in that: Based on the target intention information of each question information, combined with the historical data related to the target merchant and the target intention information of each question information, the target reply information corresponding to each question information is generated, including: For each question information, according to the target intention information of the question information, determine the target data type related to the target intention information; According to the target data type, acquiring historical data of the target data type from the first database; Based on the target intention information, obtaining a plurality of reply information adapted to the target intention information from a second database as a plurality of candidate reply information; The target answer information of the question information is determined according to the answer priorities of the plurality of candidate answer information and the historical data.

6. The method according to claim 5, characterized in that Determining target answer information for the question information according to the answer priorities of the plurality of candidate answer information and the historical data, including: According to the respective reply priorities of the plurality of candidate reply information, determining in turn whether the historical data meets the data condition contained in each candidate reply information; The candidate reply information corresponding to the data condition that matches the historical data and is closest to the historical data is used as the target reply information.

7. The method according to claim 3, characterized in that Before generating target reply information corresponding to each question information based on the target intention information of each question information and combining the historical data related to the target merchant and the target intention information of each question information, the method further includes: Based on a pre-configured complexity level of the intent information, identifying the complexity of the target intent information; If the AI ​​question-and-answer service model is adapted to the complexity of the target intent information, the AI ​​question-and-answer service model is used to obtain historical data and multiple candidate reply information related to the target intent information of the target merchant based on the target intent information.

8. The method according to any one of claims 1 to 7, characterized in that: Also includes: Displaying a hosting service page, the hosting service page including: a hosting QR code; In response to a successful scanning operation of the customer service personnel based on the hosted QR code, an intelligent customer service corresponding to the customer service personnel is determined, and the intelligent customer service is assigned to the hosted customer service queue.

9. The method according to any one of claims 1 to 7, characterized in that: Also includes: Inputting the second model prompt word into the AI ​​question-answering service model, wherein the second model prompt word is used to prompt the AI ​​question-answering service model to provide a scheduled marketing push service; Under the prompt of the second model prompt word, according to at least one target intention information corresponding to the question information of each merchant within a preset time period, a marketing plan adapted to the at least one target intention information is obtained; A marketing plan adapted to the at least one target intent information is sent to a corresponding client so that the corresponding merchant can view the marketing plan.

10. An electronic device, characterized in that: include: Memory and processor; The memory is used to store a computer program; the processor is coupled to the memory and is used to execute the computer program to implement the steps in the method according to any one of claims 1 to 9.

11. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the processor is caused to implement the steps in the method according to any one of claims 1 to 9.

12. A computer program product, characterized in that The computer program product comprises a computer program / instruction, which, when executed by a processor, enables the processor to implement the steps of any one of the methods of claims 1 to 9.

Images