Intelligent agent method and device based on intelligent language model, equipment and medium

Abstract

The invention relates to an intelligent agent method and device based on an intelligent language model, equipment and a medium, and the intelligent agent method based on the intelligent language model comprises the steps: setting corresponding interaction parameters according to the type of a communication protocol, so as to establish a data interaction channel with a target customer service system; on the basis of the intelligent language model, multiple rounds of dialogue management processes are executed on the user input text, corresponding user consultation content is generated, and the user consultation content comprises holiday and festival consultation verbal skills generated for a target area in the user input text; receiving response information returned by the target customer service system based on the user consultation content, and extracting structured field information used for optimizing the multi-round dialogue management process; generating and pushing a corresponding question-seeking verbal skill; according to the invention, the docking efficiency in a multi-platform and multi-protocol environment is effectively improved, the dependence on manual configuration is reduced, and the processing efficiency and accuracy of an intelligent customer service system in a high-concurrency scene are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of e-commerce service agents, and in particular to an intelligent agent method, device, equipment and medium based on an intelligent language model. Background Art

[0002] Currently, users on e-commerce platforms and customer service systems frequently inquire about detailed information such as product specifications, inventory status, and delivery times. Traditional manual customer service processes, which require responding to user requests one by one, struggle to efficiently respond in high-concurrency scenarios, resulting in low overall service efficiency. Furthermore, manual record-keeping can easily lead to omissions of key information or inconsistent representations, compromising the accuracy and consistency of information conveyed. Summary of the Invention

[0003] In order to solve the problem of low efficiency and accuracy of traditional manual customer service processing, the present application provides an intelligent agent method, device, equipment and medium based on an intelligent language model.

[0004] The above-mentioned invention objective of this application is achieved through the following technical solutions: An intelligent agent method based on an intelligent language model, the intelligent agent method based on an intelligent language model comprising: Identify the communication protocol type of the target customer service system, and set corresponding interaction parameters according to the communication protocol type to establish a data interaction channel with the target customer service system; Based on the intelligent language model, a multi-round dialogue management process is executed on the user input text to generate corresponding user consultation content, and the user consultation content is sent to the target customer service system through the data interaction channel. The user consultation content includes holiday consultation scripts generated for the target region in the user input text; Receiving response information returned by the target customer service system based on the user consultation content, and extracting structured field information for optimizing the multi-round dialogue management process based on the response information; Generate and push corresponding follow-up dialogue based on the response information and the structured field information; Obtain user feedback information, and determine corresponding strategy adjustment parameters based on the user feedback information, wherein the strategy adjustment parameters are used to optimize the generation process of the user consultation content and the follow-up questioning techniques.

[0005] By adopting this technical solution, the system automatically identifies the communication protocol used by the target customer service system and sets the corresponding interaction parameters based on the identified protocol, achieving flexible adaptation of data interaction channels between different customer service systems. This mechanism effectively improves the system's deployability and docking efficiency in multi-platform and multi-protocol environments, reduces reliance on manual configuration, and facilitates the high scalability and stability of the intelligent agent system.

[0006] In a preferred example, the present application may be further configured as follows: the step of identifying the communication protocol type of the target customer service system and setting corresponding interaction parameters according to the communication protocol type to establish a data interaction channel with the target customer service system includes: receiving multimodal input data provided by a user, wherein the multimodal input data includes text data, image data, and table data; Performing a parsing operation on the multimodal input data to generate a corresponding user initial demand, the parsing operation including an image recognition operation for extracting text content from image data using an optical character recognition algorithm, and a table recognition operation for extracting field information from table data using a structured parsing method; Classify the user's initial needs and generate corresponding target consultation intention types; Determining the communication protocol type of the target customer service system based on a preset technical architecture of the target customer service system and the target consultation intent type, wherein the preset technical architecture includes a RESTful API architecture and a WebSocket architecture; Corresponding interaction parameters are set according to the communication protocol type to establish a data interaction channel with the target customer service system.

[0007] By adopting this technical solution, a unified parsing mechanism for multimodal input data is introduced, enabling the system to accept user input in various formats, including text, images, and tables. It then combines optical character recognition with structured data parsing to extract initial user requirements. This is then used to classify intent using a pre-trained language model, accurately identifying the user's true inquiry intent and matching it with the appropriate communication protocol. This solution effectively improves the system's ability to handle unstructured and mixed data scenarios, enhancing the accuracy of user intent recognition and the automation of interface adaptation.

[0008] In a preferred example, the present application may be further configured as follows: in the step of executing a multi-round dialogue management process on the user input text based on the intelligent language model to generate corresponding user consultation content, the multi-round dialogue management process includes at least a semantic understanding process and a context tracking process, and the step further includes: Based on the intelligent language model, the semantic understanding process is performed on the user input text to generate the corresponding user semantic expression; A context tracking process is performed on the user semantic expression, and user consultation content for the current round is generated based on the historically extracted structured field information.

[0009] By adopting the above technical solutions and building a multi-round dialogue management mechanism based on semantic understanding and context tracking, the system can continuously maintain semantic state during user interaction, dynamically generating consultation content for the current round based on historically extracted information, and ensuring the coherence and integrity of the dialogue process. This mechanism establishes stronger dialogue logic support at the semantic control level, enabling the system to perform continuous, multi-step information confirmation for complex needs, significantly enhancing the naturalness and intelligence of interactions.

[0010] In a preferred example, the present application may be further configured as follows: after the step of performing a semantic understanding process on the user input text to generate a corresponding user semantic expression, the application may further include: detecting whether a holiday awareness trigger field appears in the user semantic expression, and if the holiday awareness trigger field does not appear, preparing to execute the step of executing the context tracking process for the user semantic expression; If a holiday awareness trigger field appears, the corresponding target region is determined based on the holiday awareness trigger field, and the statutory holiday information of the target region is matched based on the public holiday API or the local holiday database; According to the statutory holiday information, push the corresponding holiday consultation script; If it is detected that the holiday consultation script has been responded to, historical shipping data is obtained, and the corresponding logistics delay time range is calculated and automatically pushed based on the historical shipping data and the statutory holiday information.

[0011] By implementing this technical solution, a holiday-aware trigger mechanism is introduced into the semantic parsing process. This mechanism determines whether a holiday context is involved based on key fields in the user's input text. It then uses the public holiday API or local holiday database to retrieve statutory holiday information for the corresponding country or region, generating relevant scripts and providing logistics delay predictions. This mechanism significantly improves the system's information timeliness and adaptability in special scenarios such as cross-regional and holiday situations, enhancing the user experience during sensitive time windows.

[0012] In a preferred example, the present application may be further configured as follows: the step of extracting structured field information for optimizing the multi-round dialogue management process based on the response information includes: Extracting key fields from the response information; Performing logic verification on the key fields based on a local verification database; Mapping the key fields that pass the logic check to a predefined local mapping library to match the corresponding structure fields; Each of the structured fields is classified and stored to generate corresponding structured field information.

[0013] By employing this technical solution, when processing responses from the customer service system, the system extracts key fields and performs logical validation and semantic mapping operations on the local database, ultimately generating structured field information that conforms to a predefined structure. This structuring process provides a reliable data foundation for the system's subsequent semantic tracking and strategy optimization, improving the consistency and accuracy of information processing and the system's ability to continuously accumulate knowledge.

[0014] In a preferred example, the present application may be further configured as follows: the step of generating and pushing corresponding follow-up dialogue according to the response information and the structured field information includes: Determine whether the response information is complete, and if so, wait for receiving new user input text; If incomplete, comparing the response information with the structured field information to identify the corresponding target field; Calling a preset speech template and completing the speech template based on the target field; A natural language reconstruction operation is performed on the completed speech template to generate a follow-up speech that is consistent with the context style in the response information.

[0015] By employing this technical solution, when customer service replies contain missing information, the system automatically identifies inconsistencies between the replies and structured field information, invokes script templates to complete the missing fields, and uses a language generation model to perform natural language reconstruction, ensuring that the generated script is semantically consistent with the current conversation style. This follow-up script generation mechanism significantly improves the system's ability to handle incomplete responses, avoids repeated user input, and effectively ensures the integrity and intelligence of the consultation process.

[0016] The second object of the present invention is achieved through the following technical solutions: An intelligent agent device based on an intelligent language model, comprising: an identification module for identifying the communication protocol type of the target customer service system and setting corresponding interaction parameters according to the communication protocol type to establish a data interaction channel with the target customer service system; The first generation module is configured to execute a multi-turn dialogue management process on the user input text based on an intelligent language model, generate corresponding user consultation content, and send the user consultation content to the target customer service system through the data interaction channel, wherein the user consultation content includes holiday consultation scripts generated for a target region in the user input text. The receiving module is configured to receive response information returned by the target customer service system based on the user consultation content, and extract structured field information for optimizing the multi-turn dialogue management process according to the response information. The second generation module is configured to generate and push corresponding follow-up questions according to the response information and the structured field information. The determining module is configured to obtain user feedback information, determine corresponding strategy adjustment parameters according to the user feedback information, and use the strategy adjustment parameters to optimize the generation process of the user consultation content and the follow-up questions.

[0017] The above-mentioned fourth purpose of the present application is achieved through the following technical solutions: A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor executes the computer program to implement the steps of the above-mentioned intelligent agent method based on an intelligent language model.

[0018] The above-mentioned fourth purpose of the present application is achieved through the following technical solutions: A computer readable storage medium stores a computer program, and the computer program is executed by a processor to implement the steps of the above-mentioned intelligent agent method based on an intelligent language model.

[0019] In summary, the present application includes at least one of the following beneficial technical effects: This application automatically connects to different customer service platforms by identifying the communication protocol type of the target customer service system and dynamically configuring interaction parameters, improving the system's compatibility and deployment flexibility across multiple platforms. Subsequently, by executing a multi-round dialogue management process on user input text, combined with context tracking and semantic parsing mechanisms, it continuously identifies evolving user needs and generates coherent and contextually relevant user consultation content, significantly improving the accuracy and completeness of responses during conversations. It also supports generating holiday-adaptive scripts based on the target region in the user text. Using external holiday data sources or local databases, it dynamically perceives holiday factors and adjusts scripts, addressing the information lag and rigid expression issues of traditional systems in holiday contexts. After receiving the customer service system's response information, the system further extracts key fields and performs structured processing to provide semantic support for subsequent follow-up script generation and multi-round process optimization. In combination with user feedback, the system can dynamically adjust the consultation content generation strategy to continuously optimize interaction performance. By integrating the deep semantic understanding capabilities of the intelligent language model with the interaction control mechanism, the intelligent customer service system improves its processing efficiency in high-concurrency scenarios, its adaptability to complex languages, and its response accuracy and user experience under cross-regional holiday conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is a flow chart of an intelligent agent method based on an intelligent language model in one embodiment of the present application.

[0021] Figure 2 This is a flowchart for implementing step S10 in an intelligent agent method based on an intelligent language model in one embodiment of the present application; Figure 3 This is a flowchart for implementing step S20 in an intelligent agent method based on an intelligent language model in one embodiment of the present application; Figure 4 This is another implementation flowchart of step S20 in an intelligent agent method based on an intelligent language model in one embodiment of the present application; Figure 5 This is a flowchart for implementing step S30 in an intelligent agent method based on an intelligent language model in one embodiment of the present application; Figure 6 This is a flowchart for implementing step S40 in an intelligent agent method based on an intelligent language model in one embodiment of the present application; Figure 7 This is a principle block diagram of an intelligent agent device based on an intelligent language model in one embodiment of the present application; Figure 8 It is a schematic diagram of a device in one embodiment of the present application. DETAILED DESCRIPTION

[0022] The present application is further described in detail below with reference to the accompanying drawings.

[0023] In one embodiment, if Figure 1 As shown, the present application discloses an intelligent agent method based on an intelligent language model, and the intelligent agent method based on an intelligent language model includes: S10. Identify the communication protocol type of the target customer service system and set corresponding interaction parameters based on the communication protocol type to establish a data interaction channel with the target customer service system. The communication protocol type of the target customer service system refers to the data interaction mechanism adopted by the target customer service platform in its technical implementation, such as the RESTful API architecture based on the HTTP protocol or the WebSocket architecture based on persistent connections. Different communication protocols will affect the system's access logic in terms of request methods, connection persistence, and data format processing. Interaction parameters are a set of connection and communication configuration items set for the identified communication protocol, including but not limited to request header parameters, authentication information, data format specifications, and connection maintenance strategies. They are used to ensure that a stable and standardized data interaction channel can be established between the intelligent agent system and the target customer service system.

[0024] S20. Based on the intelligent language model, a multi-round dialogue management process is executed on the user input text to generate corresponding user consultation content. This user consultation content is then sent to the target customer service system via a data exchange channel. The user consultation content includes holiday consultation scripts generated for the target region in the user input text. The intelligent language model is a deep neural network model trained on large-scale corpus, capable of natural language understanding and generation. In this method, it is used to analyze the semantic features, contextual relationships, and intent content in the user input text, providing a semantic foundation for subsequent dialogue management and script generation. User consultation content refers to the natural language expression generated by the system based on the user input text, historical semantic state, and the current stage of the dialogue. It reflects the user's current specific consultation needs and may include information such as product parameter inquiries, logistics status inquiries, and holiday shipping arrangements. It is the primary information carrier for the system to actively interact with the customer service platform. Holiday consultation scripts are time-sensitive dialogue content generated by the system based on holiday recognition results when the user text mentions specific regions or time-sensitive expressions. They are used to proactively remind or supplement special arrangements such as shipping and customer service responses during holidays.

[0025] S30. Receive the response information returned by the target customer service system based on the user's consultation content, and extract structured field information for optimizing the multi-round dialogue management process based on the response information; structured field information refers to the semantic content extracted by the system from the response information returned by the customer service platform that conforms to the predefined data format or business field model. These fields are usually standardized and classified to support context judgment, missing information identification, and speech generation optimization of multi-round dialogues.

[0026] S40. Generate and push corresponding follow-up dialogue based on the response information and structured field information. Follow-up dialogue is a supplementary natural language expression generated by the system based on the current context and missing field type after the system identifies that key information is missing in the customer service response. Its purpose is to guide the customer service to further clarify the unanswered key information to maintain the coherence and semantic integrity of the dialogue process.

[0027] S50. Obtain user feedback information and determine corresponding strategy adjustment parameters based on the user feedback information. The strategy adjustment parameters are used to optimize the generation process of user consultation content and follow-up questioning techniques. User feedback information refers to the response signal implicitly or explicitly provided by the user after completing a round of interaction, which may include re-consultation behavior, rating operations, cancellation requests, etc., and is used to evaluate the accuracy and satisfaction of the content generated by the current system.

[0028] Specifically, when a user enters the question "Can we ship during the Spring Festival?", the system first identifies that the target customer service system uses the WebSocket protocol and configures communication parameters. It then uses an intelligent language model to analyze the user text and generate a user inquiry message, including information about Spring Festival delivery times, which is then sent to the customer service platform. The customer service platform responds with "We need to check inventory before shipping." The system extracts the structured field "Inventory status," but finds that it doesn't include a specific shipping time field. Therefore, it automatically generates and pushes the follow-up question, "What is the estimated shipping time during the Spring Festival?" The system also records the user's subsequent behavior and uses it to update policy parameters to optimize the performance of the generated query in similar scenarios in the future.

[0029] In one embodiment, if Figure 2 As shown, in step S10, i.e., identifying the communication protocol type of the target customer service system and setting corresponding interaction parameters according to the communication protocol type to establish a data interaction channel with the target customer service system, the steps include: S101. Receive multimodal input data provided by a user. Multimodal input data includes text data, image data, and tabular data. Multimodal input data refers to the collection of original content submitted to the system by the user in different information forms, typically including multiple types of data input, such as text, images, or structured tables. Text data can come from natural language instructions or inquiry statements directly entered by the user. Image data may be product screenshots, interface photos, or other information images that require further recognition. Tabular data is often a collection of fields with structural relationships, such as product lists, inventory tables, and shipping schedules.

[0030] S102. Perform parsing operations on multimodal input data to generate corresponding initial user requirements. Parsing operations include image recognition operations that extract text content from image data using an optical character recognition algorithm, and table recognition operations that extract field information from table data using a structured parsing method. Image recognition operations are the process of parsing image data. The system uses the deployed optical character recognition algorithm to identify readable characters contained in the image and extract potential text information for further constructing user requirement content. Table recognition operations refer to the process of performing field-level analysis and extraction on structured data tables. The system parses the semantic relationships between columns and fields based on a predefined table structure template or machine learning model, converting each information unit in the table into a field expression that can be understood by the system.

[0031] S103. Classify the user's initial request intent and generate the corresponding target consultation intent type. The user's initial request is an abstract representation of the user's true intent generated by the system based on the analysis of multimodal input. It reflects the target direction of the user's current consultation or operation and serves as the basic data source for subsequent intent classification and dialogue management. The target consultation intent type is the specific interaction intent identified by the system after classifying the user's initial request. Common types include product parameter query, inventory availability confirmation, logistics timeliness inquiry, and holiday shipping schedule determination. The determination of the intent type affects the selection of subsequent dialogue paths and the configuration of interaction methods.

[0032] S104. Determine the communication protocol type of the target customer service system based on the preset technical architecture and target consultation intent type of the target customer service system. The preset technical architecture includes the RESTful API architecture and the WebSocket architecture. The preset technical architecture refers to a set of capability tags used within the system to model and classify the communication characteristics of different customer service platforms. It represents the technical communication methods supported by various customer service systems pre-defined by the system, such as the RESTful API architecture based on the request-response mechanism or the WebSocket architecture based on the persistent connection mechanism. The system determines the communication protocol form to be adopted through the combination of this architecture information and the intent type.

[0033] S105: Setting corresponding interaction parameters according to the communication protocol type to establish a data interaction channel with the target customer service system.

[0034] Specifically, we first build an intelligent agent system based on an intelligent language model. This intelligent agent system is capable of perceiving and parsing multimodal input data provided by users, including text, images, and tables. The system performs optical character extraction (OCE) on image data through image recognition, extracting recognizable text information. It also performs structured parsing on table data through table recognition, extracting the semantic content corresponding to the fields, and then generating the user's initial requirements.

[0035] After obtaining the user's initial request, the system executes a semantic understanding process based on a trained intelligent language model, classifying the user's initial request intent and identifying the target inquiry intent type. Target inquiry intent types include predefined categories such as product specification inquiry, delivery cycle confirmation, and inventory status acquisition. Based on the recognition results, the system generates preliminary user inquiry content, which is subsequently incorporated into the multi-round dialogue management process for context tracking and optimization generation.

[0036] To automatically connect the intelligent agent system to the target customer service system, the system further obtains information about the pre-set technical architecture of the target customer service system, which includes both a RESTful API and a WebSocket architecture. Based on the relationship between the target inquiry intent type and the pre-set technical architecture, the system determines the target customer service system's communication protocol type and sets corresponding interaction parameters accordingly. These parameters include configuration items such as the request method, authentication method, and data packet structure. Finally, through the configured data interaction channel, the intelligent agent system is ready for connection with the target customer service system, providing the communication foundation for the subsequent sending of user inquiry content and receiving responses.

[0037] In one embodiment, if Figure 3 As shown, in step S20, i.e., based on the intelligent language model, a multi-round dialogue management process is executed on the user input text to generate corresponding user consultation content. The multi-round dialogue management process includes at least a semantic understanding process and a context tracking process, and the steps also include: S2011. Based on the intelligent language model, the semantic understanding process is performed on the user input text to generate the corresponding user semantic expression. The user semantic expression refers to the intermediate semantic structure result obtained by the system after completing the semantic understanding process and performing semantic analysis on the user input text. This result is not directly equivalent to the original user input, but is a structured semantic representation with grammatical classification, semantic role labeling, and intent mapping information after being processed by the intelligent language model. The user semantic expression usually contains semantic components such as query goals, constraints, time expressions, product entities, etc., which are used for logical deduction and context fusion in the subsequent dialogue control module. It is the semantic driving core that promotes the evolution of multi-round dialogues. S2012. Execute the context tracking process for the user's semantic expression, and generate the user consultation content for the current round based on the historically extracted structured field information; the context tracking process refers to the process by which the system continuously perceives and dynamically manages the historical conversation content and the current semantic state during multiple rounds of conversations. Its core goal is to maintain semantic continuity and conversation state consistency. This process does not rely solely on the current input text, but combines the user's semantic expression for the current round with the historically extracted structured field information to identify the conversation stage, detect missing information, and backtrack confirmed content, thereby generating the most appropriate user consultation content for the current round. The historically extracted structured field information serves as the context reference basis in this process, helping the system determine which fields have been confirmed and which fields are to be completed, thereby avoiding repeated questions or missing key fields when generating user consultation content.

[0038] Specifically, the state management logic integrated into the context tracking process uses a state machine modeling approach to record the questions asked by the system, user feedback, and structured field information returned by the target customer service system in each round of interaction, forming a dialogue path identifier and phased state record to support the orderly connection of cross-round content judgment and dialogue context.

[0039] In one embodiment, if Figure 4 As shown, after step S2011, that is, after the step of executing the semantic understanding process on the user input text and generating the corresponding user semantic expression, the following steps are also included: S2021. Detect whether a holiday-aware trigger field appears in the user's semantic expression. If not, prepare to execute the steps of executing the context tracking process for the user's semantic expression. Holiday-aware trigger fields refer to semantic keywords or time expression units associated with holidays or specific dates in the user's semantic expression. These typically include expressions such as holiday names, time intervals, and holiday schedules, such as "during the Spring Festival," "National Day holiday," or "Labor Day holiday." When the system identifies these fields during the semantic understanding process, it determines that the current consultation content involves a holiday context, thereby triggering the information processing process associated with holidays.

[0040] S2022. If a holiday-aware trigger field is present, the corresponding target region is determined based on the holiday-aware trigger field. Legal holiday information for the target region is matched based on the public holiday API or the local holiday database. The target region refers to the spatial location identifier at the national, regional, or city level inferred by the system based on the geographical information reflected in the holiday-aware trigger field or the geographical target implied by the user's intent. This target region serves as the key input for querying legal holiday information. The duration and schedule of holidays for the same holiday may vary across different regions, so a clear geographic reference is required to ensure accurate holiday matching. The public holiday API is an online data interface service provided to developers for obtaining legal holiday information globally or for specific countries and regions. This API returns structured data such as the corresponding holiday time, name, and type through standard network requests. The local holiday database is an offline data set pre-integrated in the local system that contains holiday schedules for multiple regions and supports holiday data queries in offline environments. These two data sources can be selected based on the system deployment environment to ensure the coverage and timeliness of holiday information.

[0041] S2023. Push corresponding holiday consultation scripts based on statutory holiday information. Holiday consultation scripts are natural language dialogues automatically generated by the system based on statutory holiday information in the target region after the system recognizes the user's holiday-related intent. They are primarily used to inform users of possible service differences during holidays, such as shipping adjustments and changes in customer service responses, to help users form reasonable expectations and improve the contextual adaptability of system responses. The language style and content structure of this script should be consistent with the current conversation context to avoid abrupt insertion of information that affects the user experience.

[0042] S2024. If it is detected that the holiday consultation script has been responded to, the historical shipping data is obtained, and the corresponding logistics delay time range is calculated and automatically pushed based on the historical shipping data and statutory holiday information; historical shipping data refers to the logistics processing records of goods in different time periods, holiday cycles or specific regions pre-stored in the system. The data content includes shipping processing time, transportation time, distribution fluctuation range, etc. This data is used to evaluate the logistics performance during similar holidays, thereby providing the current user with a more valuable reference for shipping time estimates. The logistics delay time range is the shipping time variation range calculated by the system based on the holiday information and historical shipping data of the target area. It is usually presented in the form of "delay of X to Y working days" and is used to give active prompts during the user consultation process to help them evaluate the potential impact between order placement and delivery time.

[0043] Specifically, the system introduces a holiday perception module, which is used to determine whether the current consultation involves a holiday scenario based on whether a holiday perception trigger field appears in the user's semantic expression.

[0044] When the system detects the presence of a holiday-aware trigger field in a user's semantic expression, it first determines the target region based on the geographical information in that field. It then calls pre-defined data interfaces, including the public holiday API or local holiday database, to query the target region's statutory holiday information in real time. The holiday-aware module recognizes structured holiday data, including holiday names, start and end times, and holiday types, and incorporates it into the current semantic context.

[0045] After acquiring holiday information, the system automatically adjusts the original user inquiry content based on the contextual state of the current multi-round dialogue management process, generating holiday inquiry scripts. These scripts are customized based on the language habits of the target region and the user's current inquiry intent. For example, if the original inquiry content is "When will the product be shipped?", the system can dynamically generate the holiday inquiry script "Does your company arrange shipments during the Spring Festival?" based on the target region's rules for suspending shipments during holidays, and send it to the target customer service system through the data exchange channel.

[0046] When the system detects that the holiday inquiry has been addressed by the customer service system, it further calls the historical shipping data module to extract historical logistics processing records that match the current product category, target region, and holiday period. This record includes parameters such as the shipping start time and actual delivery time under similar holiday scenarios. Based on this data, the system establishes a delay prediction model. Combining statutory holiday information with actual delivery time, it calculates the logistics delay range for the current scenario and automatically generates notification content in natural language and pushes it to the user.

[0047] To enhance cross-border multilingual interaction adaptability, the system leverages an integrated multilingual generation model when generating holiday consultation scripts and logistics delay alerts, translating the generated results into more natural and culturally appropriate expressions for the user's language environment. For example, when the target region is the UK and the user speaks English, the system reconstructs the expression "holiday" as "festive season," making the holiday script more consistent with local user language preferences and improving the intelligent agent system's accuracy and interactive experience in multilingual and multicultural scenarios.

[0048] In one embodiment, if Figure 5 As shown, in step S30, i.e., the step of extracting structured field information for optimizing the multi-round dialogue management process based on the response information, the following steps are included: S301. Extract key fields from the response information. Key fields are core data segments identified from the response information returned by the target customer service system that are directly related to the user's current inquiry intent. These fields typically appear in natural language but carry clear product parameters, status information, or timeliness data, such as "shipping in 3-5 business days," "out of stock," or "model X300." Extracting key fields relies on an intelligent language model to perform semantic decomposition and intent annotation on the response sentence, ensuring that semantically valid content is identified from unstructured text and entered into subsequent processing.

[0049] S302. Perform logical verification of key fields based on a local verification database. Key fields are core data segments identified from the response information returned by the target customer service system that are directly related to the user's current inquiry intent. These fields typically appear in natural language but carry clear product parameters, status information, or timeliness data, such as "Shipping in 3-5 business days," "In stock," or "Model X300." Extraction of key fields relies on an intelligent language model to perform semantic decomposition and intent annotation on the response statement, ensuring that semantically valid content is identified from unstructured text and entered into subsequent processing.

[0050] S303. Map the key fields that pass the logical validation to a predefined local mapping library to match the corresponding structured fields. The local mapping library is a set of structured field mapping relationships preset in the system, which is used to convert the key fields that pass the logical validation into a standardized structured field format for use in multi-round dialogue management processes. The mapping library establishes mapping rules from natural language descriptions to structured field labels. For example, "3-5 working days" is mapped to the field name "estimatedDeliveryTime" and "X300" is mapped to the field name "productModel", allowing the system to consistently use a unified field representation across different conversations.

[0051] S304. Classify and store each structure field to generate corresponding structured field information; classified storage means that the system groups and stores multiple mapped structure fields according to attribute type, field purpose, or business module. For example, the shipping information field, product specification field, and customer service response status field are stored in different structured information partitions to improve field call efficiency and context management capabilities.

[0052] In one embodiment, if Figure 6 As shown, in step S40, i.e., the step of generating and pushing corresponding follow-up dialogue according to the response information and the structured field information, the following steps are included: S401, determine whether the response information is complete, if complete, wait to receive new user input text; S402. If incomplete, the system compares the response information with the structured field information to identify the corresponding target fields. Target fields are key fields that are not fully addressed or whose semantic expression is incomplete in the response information. These fields are identified by comparing and analyzing the response content returned by the current customer service system with the stored structured field information. These fields typically correspond to the user's core inquiry points in the current round, such as product model, delivery time, price, or inventory. If the system finds that the response content does not clearly address these fields or only provides vague responses such as "Please wait" or "Need to confirm", it will determine that the response is incomplete and locate the specific target fields.

[0053] S403. Call the preset speech template and complete the speech template based on the target field. The speech template is a set of semi-structured language templates preset in the system, which are used to generate standard questions or supplementary dialogue content that conforms to the contextual logic. The template usually contains field placeholders that can be dynamically replaced, such as "What is the [field name]?", "Can you confirm the [field name]?", etc. After the system identifies the target field, it embeds it into the corresponding template as a fill-in value to form the preliminary completed speech content. The template library covers a variety of common consulting semantic scenarios such as products, services, shipments, prices, etc., and can adapt to different types of user intentions and customer service response styles.

[0054] S404: Perform natural language reconstruction on the completed speech template to generate follow-up speech that aligns with the contextual style of the response information. Natural language reconstruction involves the system using an intelligent language model to adjust the language style and semantic integrity of the initial speech template after generating it, making it more consistent with the current conversation context and possessing stronger human expression characteristics. This operation not only replaces vocabulary and optimizes word order, but also considers factors such as user language habits and the customer service response style, thereby generating follow-up speech that aligns with the contextual tone, is clearly expressed, and has natural transitions. The generated speech avoids mechanical template-based expression, enhancing user acceptance and interactive experience.

[0055] It should be understood that the size of the serial numbers of the steps in the above embodiments does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of this application.

[0056] In one embodiment, an intelligent agent device based on an intelligent language model is provided, and the intelligent agent device based on an intelligent language model corresponds one-to-one to an intelligent agent method based on an intelligent language model in the above embodiment. Figure 7As shown, the intelligent agent device based on the intelligent language model includes a recognition module, a first generation module, a receiving module, a second generation module and a determination module. The functional modules are described in detail as follows: An intelligent agent device based on an intelligent language model, comprising: an identification module for identifying the communication protocol type of the target customer service system and setting corresponding interaction parameters according to the communication protocol type to establish a data interaction channel with the target customer service system; a first generation module configured to execute a multi-round dialogue management process on the user input text based on an intelligent language model, generate corresponding user consultation content, and send the user consultation content to the target customer service system via the data interaction channel, wherein the user consultation content includes holiday consultation scripts generated for the target region in the user input text; a receiving module, configured to receive response information returned by the target customer service system based on the user consultation content, and extract structured field information for optimizing the multi-round dialogue management process based on the response information; A second generation module is configured to generate and push corresponding follow-up dialogue based on the response information and the structured field information; A determination module is used to obtain user feedback information and determine corresponding strategy adjustment parameters based on the user feedback information. The strategy adjustment parameters are used to optimize the generation process of the user consultation content and the follow-up questioning language.

[0057] Optionally, the identification module includes: a receiving unit, configured to receive multimodal input data provided by a user, wherein the multimodal input data includes text data, image data, and table data; a first execution unit, configured to perform a parsing operation on the multimodal input data to generate a corresponding user initial demand, the parsing operation including an image recognition operation for extracting text content from image data using an optical character recognition algorithm, and a table recognition operation for extracting field information from table data using a structured parsing method; A first generating unit is configured to classify the user's initial needs based on an intelligent language model and generate a corresponding target consultation intention type; a determining unit, configured to determine a communication protocol type of the target customer service system according to a preset technical architecture of the target customer service system and the target consultation intention type, wherein the preset technical architecture includes a RESTful API architecture and a WebSocket architecture; An establishing unit, configured to set corresponding interaction parameters according to the communication protocol type to establish a data interaction channel with a target customer service system; Optionally, in the first generation module, the multi-round dialogue management process includes at least a semantic understanding process and a context tracking process, and the first generation module further includes: The second generation unit is used to perform a semantic understanding process on the user input text based on the intelligent language model to generate a corresponding user semantic expression; A second execution unit is configured to execute a context tracking process on the user semantic expression and generate user consultation content for the current round based on the historically extracted structured field information; Optionally, the first generating module further includes: a detection unit, configured to detect whether a holiday awareness trigger field appears in the user semantic expression, and if the holiday awareness trigger field does not appear, prepare to execute the step of executing the context tracking process for the user semantic expression; A first matching unit is configured to, if a holiday awareness trigger field appears, determine a corresponding target region according to the holiday awareness trigger field, and match statutory holiday information of the target region based on a public holiday API or a local holiday database; A push unit, configured to push corresponding holiday consultation scripts according to the statutory holiday information; a calculation unit, configured to obtain historical shipment data if it is detected that the holiday consultation script has been responded to, and calculate and automatically push a corresponding logistics delay time range based on the historical shipment data and the statutory holiday information; Optionally, the receiving module includes: An extraction unit, configured to extract key fields from the response information; A verification unit, configured to perform a logic verification on the key fields based on a local verification database; A second matching unit is used to map the key fields that pass the logic check to a predefined local mapping library to match corresponding structure fields; a third generating unit, configured to classify and store each of the structure fields, and generate corresponding structured field information; Optionally, the second generating module includes: a judging unit, configured to judge whether the response information is complete, and if so, wait for receiving new user input text; a unit for comparing the response information with the structured field information to identify a corresponding target field if the response information is incomplete; A calling unit, configured to call a preset speech template and complete the speech template based on the target field; The fourth generating unit is used to perform a natural language reconstruction operation on the completed speech template to generate a follow-up speech that is consistent with the context style in the response information.

[0058] For the specific definition of an intelligent agent device based on an intelligent language model, please refer to the definition of an intelligent agent method based on an intelligent language model above, which will not be repeated here. The various modules in the above-mentioned intelligent agent device based on an intelligent language model can be implemented in whole or in part by software, hardware and a combination thereof. The above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory of the computer device in the form of software, so that the processor can call and execute the operations corresponding to the above modules.

[0059] In one embodiment, a computer device is provided. The computer device may be a server, and its internal structure diagram may be as follows: Figure 8 As shown. The computer device includes a processor, memory, network interface, and database connected via a system bus. The processor of the computer device is used to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of the operating system and computer program in the non-volatile storage medium. The network interface of the computer device is used to communicate with an external terminal via a network connection. When the computer program is executed by the processor, an intelligent agent method based on an intelligent language model is implemented.

[0060] In one embodiment, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, the following steps are performed: S10, identifying the communication protocol type of the target customer service system, and setting corresponding interaction parameters according to the communication protocol type to establish a data interaction channel with the target customer service system; S20: Based on the intelligent language model, a multi-round dialogue management process is executed on the user input text to generate corresponding user consultation content, and the user consultation content is sent to the target customer service system through the data interaction channel, where the user consultation content includes holiday consultation scripts generated for the target region in the user input text; S30: receiving response information returned by the target customer service system based on the user consultation content, and extracting structured field information for optimizing the multi-round dialogue management process based on the response information; S40: Generate and push corresponding follow-up dialogue based on the response information and the structured field information; S50: Obtain user feedback information, and determine corresponding strategy adjustment parameters based on the user feedback information, wherein the strategy adjustment parameters are used to optimize the generation process of the user consultation content and the follow-up questioning techniques; In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the following steps are implemented: S10, identifying the communication protocol type of the target customer service system, and setting corresponding interaction parameters according to the communication protocol type to establish a data interaction channel with the target customer service system; S20: Based on the intelligent language model, execute a multi-round dialogue management process on the user input text to generate corresponding user consultation content, and send the user consultation content to the target customer service system through the data interaction channel, wherein the user consultation content includes holiday consultation scripts generated for the target region in the user input text; S30: receiving response information returned by the target customer service system based on the user consultation content, and extracting structured field information for optimizing the multi-round dialogue management process based on the response information; S40: Generate and push corresponding follow-up dialogue based on the response information and the structured field information; S50: Obtain user feedback information, and determine corresponding strategy adjustment parameters based on the user feedback information, wherein the strategy adjustment parameters are used to optimize the generation process of the user consultation content and the follow-up questioning techniques; Those skilled in the art will understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program. The computer program can be stored in a non-volatile computer-readable storage medium. When the computer program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, any reference to memory, storage, database or other media used in the embodiments provided in this application may include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM) or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in many forms such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), Synchronous Link DRAM (SLDRAM), Rambus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM), etc.

[0061] Those skilled in the art will clearly understand that for the sake of convenience and brevity of description, only the division of the above-mentioned functional units and modules is used as an example. In actual applications, the above-mentioned functions can be distributed and completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above.

[0062] The above-described embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. These modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the various embodiments of the present application, and should all be included in the scope of protection of the present application.

Claims

1. An intelligent agent method based on an intelligent language model, characterized in that: The intelligent agent method based on the intelligent language model includes: Identify the communication protocol type of the target customer service system, and set corresponding interaction parameters according to the communication protocol type to establish a data interaction channel with the target customer service system; Based on the intelligent language model, a multi-round dialogue management process is executed on the user input text to generate corresponding user consultation content, and the user consultation content is sent to the target customer service system through the data interaction channel. The user consultation content includes holiday consultation scripts generated for the target region in the user input text; Receiving response information returned by the target customer service system based on the user consultation content, and extracting structured field information for optimizing the multi-round dialogue management process based on the response information; Generate and push corresponding follow-up dialogue based on the response information and the structured field information; Obtain user feedback information, and determine corresponding strategy adjustment parameters based on the user feedback information, wherein the strategy adjustment parameters are used to optimize the generation process of the user consultation content and the follow-up questioning techniques.

2. The intelligent agent method based on the intelligent language model according to claim 1, characterized in that: The step of identifying the communication protocol type of the target customer service system and setting corresponding interaction parameters according to the communication protocol type to establish a data interaction channel with the target customer service system includes: receiving multimodal input data provided by a user, wherein the multimodal input data includes text data, image data, and table data; Performing a parsing operation on the multimodal input data to generate a corresponding user initial demand, the parsing operation including an image recognition operation for extracting text content from image data using an optical character recognition algorithm, and a table recognition operation for extracting field information from table data using a structured parsing method; Classify the user's initial needs and generate corresponding target consultation intention types; Determining the communication protocol type of the target customer service system based on a preset technical architecture of the target customer service system and the target consultation intent type, wherein the preset technical architecture includes a RESTful API architecture and a WebSocket architecture; Corresponding interaction parameters are set according to the communication protocol type to establish a data interaction channel with the target customer service system.

3. The intelligent agent method based on the intelligent language model according to claim 1, characterized in that: In the step of executing a multi-round dialogue management process on the user input text based on the intelligent language model to generate corresponding user consultation content, the multi-round dialogue management process includes at least a semantic understanding process and a context tracking process, and the step further includes: Based on the intelligent language model, the semantic understanding process is performed on the user input text to generate the corresponding user semantic expression; A context tracking process is performed on the user semantic expression, and user consultation content for the current round is generated based on the historically extracted structured field information.

4. The intelligent agent method based on the intelligent language model according to claim 1, characterized in that: After the step of executing the semantic understanding process on the user input text to generate the corresponding user semantic expression, the method further includes: detecting whether a holiday awareness trigger field appears in the user semantic expression, and if the holiday awareness trigger field does not appear, preparing to execute the step of executing the context tracking process for the user semantic expression; If a holiday awareness trigger field appears, the corresponding target region is determined based on the holiday awareness trigger field, and the statutory holiday information of the target region is matched based on the public holiday API or the local holiday database; According to the statutory holiday information, push corresponding holiday consultation scripts; If it is detected that the holiday consultation script has been responded to, historical shipping data is obtained, and the corresponding logistics delay time range is calculated and automatically pushed based on the historical shipping data and the statutory holiday information.

5. The intelligent agent method based on the intelligent language model according to claim 1, characterized in that: The step of extracting structured field information for optimizing the multi-round dialogue management process based on the response information includes: Extracting key fields from the response information; Performing logical verification on the key fields based on a local verification database; Mapping the key fields that pass the logic check to a predefined local mapping library to match the corresponding structure fields; Each of the structured fields is classified and stored to generate corresponding structured field information.

6. The intelligent agent method based on the intelligent language model according to claim 1, characterized in that: The step of generating and pushing corresponding follow-up dialogue according to the response information and the structured field information includes: Determine whether the response information is complete, and if so, wait for receiving new user input text; If incomplete, comparing the response information with the structured field information to identify the corresponding target field; Calling a preset speech template and completing the speech template based on the target field; A natural language reconstruction operation is performed on the completed speech template to generate a follow-up speech that is consistent with the context style in the response information.

7. An intelligent agent device based on an intelligent language model, characterized in that: The intelligent agent device based on the intelligent language model includes: an identification module for identifying the communication protocol type of the target customer service system and setting corresponding interaction parameters according to the communication protocol type to establish a data interaction channel with the target customer service system; a first generation module configured to execute a multi-round dialogue management process on the user input text based on an intelligent language model, generate corresponding user consultation content, and send the user consultation content to the target customer service system via the data interaction channel, wherein the user consultation content includes holiday consultation scripts generated for the target region in the user input text; a receiving module, configured to receive response information returned by the target customer service system based on the user consultation content, and extract structured field information for optimizing the multi-round dialogue management process based on the response information; A second generation module is configured to generate and push corresponding follow-up dialogue based on the response information and the structured field information; A determination module is used to obtain user feedback information and determine corresponding strategy adjustment parameters based on the user feedback information. The strategy adjustment parameters are used to optimize the generation process of the user consultation content and the follow-up questioning language.

8. The intelligent agent device based on the intelligent language model according to claim 7, characterized in that: The identification module includes: a receiving unit, configured to receive multimodal input data provided by a user, wherein the multimodal input data includes text data, image data, and table data; a first execution unit, configured to perform a parsing operation on the multimodal input data to generate a corresponding user initial demand, the parsing operation including an image recognition operation for extracting text content from image data using an optical character recognition algorithm, and a table recognition operation for extracting field information from table data using a structured parsing method; A first generating unit is configured to classify the user's initial needs based on an intelligent language model and generate a corresponding target consultation intention type; a determining unit, configured to determine a communication protocol type of the target customer service system according to a preset technical architecture of the target customer service system and the target consultation intention type, wherein the preset technical architecture includes a RESTful API architecture and a WebSocket architecture; The establishing unit is used to set corresponding interaction parameters according to the communication protocol type to establish a data interaction channel with the target customer service system.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the steps of the intelligent agent method based on the intelligent language model as described in any one of claims 1 to 6 are implemented.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the steps of the intelligent agent method based on the intelligent language model as described in any one of claims 1 to 6 are implemented.

Images