Business handling method and device based on large model and multi-agent cooperation

Abstract

The invention discloses a business handling method and device based on large model and multi-agent cooperation, and relates to the technical field of large models. The method comprises the steps of obtaining dialogue information input by a user in a business handling process; performing intention recognition on dialogue information input by a user by using an intention routing agent to obtain an intention recognition result, and determining a target business service agent matched with the intention recognition result, at least one business workflow being integrated in the target business service agent, business workflows suitable for different business scenes are integrated in different business service intelligent bodies; and calling a corresponding business workflow by using the target business service agent to perform real-time interaction with the business system to obtain a business data processing result, filling a variable corresponding to the preset dialogue template by using the business data processing result, and generating a business handling response verbal skill to reply the user. According to the application, the user can directly realize service handling through the dialogue system, and large model illusion is prevented.

Description

Business processing methods and devices based on large models and multi-agent collaboration Technical Field

[0001] This application relates to the field of large model technology, and in particular to a business processing method and apparatus based on large models and multi-agent collaboration. Background Technology

[0002] A bank's customer service dialogue system is a system that utilizes artificial intelligence technologies, such as natural language processing and machine learning, to provide automated and intelligent services to customers. It can answer inquiries, provide 24 / 7 service, and improve customer experience and efficiency.

[0003] Currently, customer service dialogue systems typically construct response information by retrieving knowledge from a knowledge base. However, the knowledge base stores static knowledge, which is only applicable to scenarios such as customer inquiries about overdue payments or billing cycles. It cannot directly process certain business transactions, such as helping customers redeem coupons or gifts. Summary of the Invention

[0004] In view of this, this application provides a business processing method and apparatus based on a large model and multi-agent collaboration, which mainly enables users to directly process business through a dialogue system, thereby improving the user experience.

[0005] According to a first aspect of this application, a business processing method based on a large model and multi-agent collaboration is provided. The method includes: acquiring dialogue information input by a user during the business processing process; using an intent routing agent to perform intent recognition on the user-input dialogue information, obtaining an intent recognition result, and determining a target business service agent matching the intent recognition result, wherein the target business service agent integrates at least one business workflow, and different business service agents integrate business workflows applicable to different business scenarios; using the target business service agent to invoke the corresponding business workflow to interact with the business system in real time, obtaining business data processing results, and using the business data processing results to fill variables corresponding to a preset dialogue template to generate a business processing response script to reply to the user.

[0006] According to a second aspect of this application, a business processing device based on a large model and multi-agent collaboration is provided. The device includes: an acquisition unit for acquiring dialogue information input by a user during the business processing process; an intent recognition unit for using an intent routing agent to perform intent recognition on the dialogue information input by the user, obtaining an intent recognition result, and determining a target business service agent matching the intent recognition result, wherein the target business service agent integrates at least one business workflow, and different business service agents integrate business workflows suitable for different business scenarios; and a business processing unit for using the target business service agent to call the corresponding business workflow to interact with the business system in real time, obtaining business data processing results, and using the business data processing results to fill variables corresponding to a preset dialogue template to generate a business processing response script to reply to the user.

[0007] According to a third aspect of this application, a storage medium is provided on which a computer program is stored, which, when executed by a processor, implements the above-described business processing method based on a large model and multi-agent collaboration.

[0008] According to a fourth aspect of this application, an electronic device is provided, including a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, wherein the processor executes the program to implement the above-described business processing method based on a large model and multi-agent collaboration.

[0009] By employing the aforementioned technical solution, this application provides a business processing method and apparatus based on a large model and multi-agent collaboration. Compared to existing technologies that retrieve knowledge by searching a knowledge base, this method utilizes an intent routing agent to identify the intent of the user's input dialogue information, obtains the intent identification result, and determines the target business service agent that matches the intent identification result. Then, the target business service agent invokes the corresponding business workflow to interact with the business system in real time, obtaining the business data processing result. Based on the business data processing result and a preset dialogue template, a business processing response is generated to reply to the user. Therefore, this application, by using an intent routing agent to identify the user's dialogue information, can accurately route user requests to the appropriate business service agent. Simultaneously, by using the business service agent to invoke the corresponding business workflow to interact with the business system in real time, it can obtain dynamic business data processing results, thereby realizing business processing. Furthermore, after obtaining the business data processing result, this application uses a standard dialogue template to construct the response, avoiding the large model illusion problem.

[0010] The above description is only an overview of the technical solution of this application. In order to better understand the technical means of this application and to implement it in accordance with the contents of the specification, and to make the above and other objects, features and advantages of this application more obvious and understandable, the following are specific embodiments of this application. Attached Figure Description

[0011] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments of this application and, together with their descriptions, serve to explain this application and do not constitute an undue limitation thereof. In the drawings: Figure 1 shows a schematic flowchart of a business processing method based on a large model and multi-agent collaboration provided by an embodiment of this application; Figure 2 shows an overall schematic flowchart of business processing provided by an embodiment of this application; Figure 3 shows a schematic flowchart of a business workflow provided by an embodiment of this application; Figure 4 shows a schematic structural diagram of a business processing device based on a large model and multi-agent collaboration provided by an embodiment of this application. Detailed Implementation

[0012] The present application will be described in detail below with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in the embodiments of the present application can be combined with each other.

[0013] The existing technical knowledge base stores static knowledge, which can only be used for fixed domain knowledge Q&A scenarios such as customer queries about what is overdue or what is a billing cycle. However, it cannot be used for some business processing scenarios, such as helping customers redeem a coupon or gift.

[0014] To address the aforementioned issues, this invention provides a business processing method based on a large model and multi-agent collaboration, as shown in Figure 1. The method includes: Step 101, obtaining dialogue information input by the user during the business processing process.

[0015] This business specifically includes banking services, securities business, and insurance business.

[0016] This invention is primarily applicable to scenarios where users can conduct business transactions through a dialogue-based question-and-answer format. The executing entity of this invention is a device or equipment capable of conducting business transactions for users through a dialogue-based question-and-answer format.

[0017] The overall dialogue service processing flow provided by this embodiment of the invention is shown in Figure 2. This flow mainly involves four core modules, namely the dialogue window, the intent routing agent, the business service agent, and the back-end business system. The dialog window serves as the front-end interface for user interaction, supporting multi-turn dialogue interactions. It receives user input in real time and quickly responds to and returns corresponding business processing responses based on the data processing results of the business system. The intent routing agent, relying on an embedded large model, performs deep semantic analysis and intent recognition on the user input dialogue information. Based on preset intent classification rules, it accurately routes user requests to the appropriate business service agent, achieving efficient service distribution. The business service agent undertakes the core responsibility of specific business processing. It integrates one, two, or more business workflows, each corresponding to a specific scenario's business processing flow. The business workflow supports flexible process orchestration, allowing data interaction nodes to be orchestrated within the business processing flow. Through remote data communication mechanisms, it achieves cross-system data collaboration and processing. The back-end business system, as the business data processing hub, establishes a two-way communication link with the business workflow. The business system performs in-depth data processing according to business needs and promptly sends back the business data processing results, ensuring the integrity and accuracy of the business processing flow.

[0018] This invention not only provides users with relevant business processing services through dialogue and question-and-answer formats, but also constructs reliable responses using script templates and data interaction results. Compared to existing technologies that directly generate responses using large models, this avoids the drawbacks of uncontrollable responses from large models. After a user sends their dialogue information, the intent routing agent first identifies the intent and undertakes the business routing function, that is, it identifies the business the user wants to process and triggers the corresponding business service agent. The business service agent integrates a business workflow. During business processing, the large model nodes within the business workflow extract business processing elements from the user's dialogue, such as time and card number, through parameter extraction capabilities. Then, it calls the backend business system to perform data interaction through the business workflow configuration and orchestration data interaction adjustment points to complete the business query or processing. Finally, based on the data interaction results, it fills the corresponding variables with a preset dialogue template, assembles the response script, and generates a reliable business processing response script to reply to the user. This forms a closed-loop service system of "user dialogue - intent identification - parameter extraction - workflow orchestration - business processing - script assembly".

[0019] In this embodiment of the invention, when a user conducts business, they can input dialogue information through a front-end dialogue portal. The customer service dialogue system can obtain this dialogue information, and the user can engage in multiple rounds of dialogue with the dialogue system through this front-end dialogue portal.

[0020] Step 102: Use the intent routing agent to perform intent recognition on the dialogue information input by the user, obtain the intent recognition result, and determine the target business service agent that matches the intent recognition result.

[0021] The target business service agent integrates at least one business workflow, and different business service agents integrate business workflows applicable to different business scenarios. The intent routing agent embeds a large intent recognition model, which can be an LLM model, a BERT model, etc.

[0022] In this embodiment of the invention, when using the intent recognition big model to perform intent recognition on the dialogue information input by the user, step 102 specifically includes: using the intent recognition big model and the preset prompt word library to perform intent recognition on the dialogue information input by the user, and obtaining intent recognition results.

[0023] This invention analyzes user-input dialogue information using an intent recognition model to determine the direction of the business process and generate multiple sub-branches of the business processing process, as shown in Figure 3.

[0024] In the specific process of intent recognition, the intent recognition big model is used to perform preliminary intent recognition on the dialogue information input by the user to obtain the confidence level of the dialogue information having different intents; if the confidence level does not meet the preset requirements, it is determined that the intent of the dialogue information is unclear, and the dialogue information is matched with prompt words using a preset prompt word library, and the intent recognition result is determined based on the prompt word matching result.

[0025] The preset requirements can be set according to actual business needs.

[0026] For example, when the confidence level is below a preset threshold (e.g., 0.8), the intent of the dialogue information is determined to be unclear. In this case, the intent recognition result can be determined by combining the preset prompt word library. As another example, if there are multiple candidate intents with similar confidence levels, the intent recognition result can also be determined by combining the preset prompt word library.

[0027] The intent routing agent in this embodiment of the invention leverages the powerful understanding capabilities of a large model and a pre-configured prompt vocabulary to enable the intent recognition model to accurately grasp the deeper business intent behind the user's natural language. Compared to traditional methods, this embodiment of the invention eliminates the need for cumbersome intent data orchestration, configuration, training, and maintenance, offering greater flexibility and intelligence.

[0028] Step 103: Utilize the target business service agent to invoke the corresponding business workflow and interact with the business system in real time to obtain the business data processing results. Then, use the business data processing results to fill the variables corresponding to the preset dialogue template to generate a business processing response script to reply to the user.

[0029] The target business service intelligence body embeds a large business processing model, specifically an LLM (Local Management Model). Different business workflow branches correspond to different preset dialogue templates. Different business workflows are applicable to different business scenarios, and the corresponding business workflows are those that match the user's intent.

[0030] In this embodiment of the invention, in order to obtain the business data processing result, step 103 specifically includes: using the business processing big model to identify and extract key business parameters from the dialogue information input by the user; and based on the key business parameters, performing real-time data interaction with the business system to obtain the business data processing result.

[0031] When interacting with business systems in real time, the communication methods supported by the business systems are first determined. Then, based on the key business parameters and the interface documents provided by the business systems, a business processing request containing the key business parameters is constructed and sent to the business systems based on the communication methods. Finally, the business data processing results fed back by the business systems based on the business processing request are received.

[0032] When generating a business processing response script based on the business data processing results and a preset dialogue template, the process first identifies placeholders in the preset dialogue template. Then, based on these placeholders, it queries a preset placeholder-to-business data key name mapping table to determine the corresponding business data key name. Finally, it replaces the value corresponding to the business data key name in the business data processing results with the value in the preset dialogue template to generate the business processing response script to reply to the user. The preset placeholder-to-business data key name mapping table records the mapping relationship between placeholders in the preset dialogue template and business data key names.

[0033] As shown in business branch 1 of Figure 3, the LLM model is first used to identify and extract key business parameters from the user's dialogue information, such as card number and date. Then, based on these key business parameters, a business processing request is generated and sent to the business system for data interaction, as shown in business systems 1-4 in Figure 3. After receiving the business processing request, the business system processes the data and feeds back the processing results to the business service agent. The business service agent then assembles the processing results with a corresponding preset dialogue template to obtain a reliable business processing response. At this point, the sub-process ends and the result is output.

[0034] In this embodiment of the invention, during business processing, key business parameters may not be extracted, and interaction with the backend business system can be performed directly. Based on this, the method further includes: enabling the data interaction node of the business workflow where the target business service agent resides to directly interact with the business system in real time to obtain the business data processing results.

[0035] As shown in business branch 2 in Figure 3, the data interaction node of the business workflow where the business service agent is located directly interacts with the back-end business system to obtain the business data processing results. Then, the business service agent assembles the business data processing results with the corresponding preset dialogue template to obtain a reliable business processing response script. At this time, the sub-process ends and the result is output.

[0036] The dialogue system in this embodiment of the invention constructs a collaborative architecture of intent routing agents and business service agents. Leveraging powerful semantic analysis capabilities, the intent routing agent can accurately identify user needs and complete business distribution, thereby ensuring efficient matching between user requests and business scenarios. The multiple business service agents in this embodiment each perform their specific duties, focusing on in-depth processing and process execution of their respective businesses. The clear division of labor and efficient collaboration among the business service agents provide users with a one-stop service experience covering the entire process.

[0037] In the workflow business branch of this invention embodiment, the "Business Processing Response Script Construction Device" node adopts a template configuration mechanism. It obtains business data processing results from the data interaction node, such as monetary data and processing result data, and then fills them into a pre-configured dialogue template to generate a reliable business processing response script. Compared with directly generating response scripts using a large model, the above method of this invention embodiment can avoid the drawbacks of the limitations of large models' professional capabilities and the uncontrollable generation results. In the overall technical solution of this invention embodiment, intent recognition and parameter extraction are first achieved using a large model, and then a fixed dialogue template is used in the user-facing response stage. This approach not only improves the intelligent interactive experience during the dialogue process but also ensures the compliance and credibility of the response results.

[0038] To make the technical solutions of the embodiments of the present invention clearer, the business processing procedure will be described in detail using the following business scenario as an example.

[0039] First, an intent routing agent and a business service agent are built on the Dify platform. Each business service agent integrates one, two, or more business workflows. The intent routing agent identifies the user's intent to process a business or check accounts through an intent recognition model. Users can engage in multi-turn dialogues with the agent "Xiaoxin Guanjia" through a front-end dialogue entry. After the intent routing agent in the background recognizes the intent, it calls the corresponding business workflow to complete the processing of the corresponding business (such as account checking, phone bill recharge, installment payment, gift redemption, coupon redemption, etc.). For example, if a user inputs the dialogue message "Help me check my spending records for last month that exceed 100 yuan", the intent routing agent recognizes that the user's intent is to check accounts, and thus distributes the dialogue request to the account checking agent. The account checking agent calls the bill details workflow to query the corresponding data, that is, extracts 100 yuan as the key business parameter from the dialogue message, requests the interface to find bill details data that meet the conditions, and then calls the preset dialogue template to generate the corresponding business processing response script for querying bill details. This enables the full application of the large model, thereby providing a smoother, more flexible, and intelligent business processing dialogue experience.

[0040] This invention provides a business processing method based on a large model and multi-agent collaboration. By using an intent routing agent to identify the user's dialogue information, the method accurately routes user requests to the appropriate business service agent. Simultaneously, the method utilizes the business service agent to invoke the corresponding business workflow and interact with the business system in real time, obtaining dynamic business data processing results to facilitate business processing. Furthermore, after obtaining the business data processing results, the method uses a standard dialogue template to construct a response, avoiding the large model illusion problem. Moreover, this invention meets the financial industry's demand for highly controllable and reliable dialogue content, while also enabling the application of the large model during business processing dialogues, fully leveraging its flexible understanding capabilities to enhance the user's business processing dialogue experience.

[0041] Furthermore, as a specific implementation of the method shown in Figure 1, this embodiment provides a business processing device based on a large model and multi-agent collaboration, as shown in Figure 4. The device includes: an acquisition unit 31, an intent recognition unit 32, and a business processing unit 33.

[0042] The acquisition unit 31 can be used to acquire dialogue information entered by the user during the business processing.

[0043] The intent recognition unit 32 can be used to perform intent recognition on the dialogue information input by the user using the intent routing agent, obtain the intent recognition result, and determine the target business service agent that matches the intent recognition result. The target business service agent integrates at least one business workflow, and different business service agents integrate business workflows suitable for different business scenarios.

[0044] The business processing unit 33 can be used to call the corresponding business workflow and interact with the business system in real time using the target business service agent to obtain business data processing results, and use the business data processing results to fill the variables corresponding to the preset dialogue template to generate business processing response scripts to reply to the user.

[0045] In some embodiments, the intent routing intelligence body embeds an intent recognition model, and the intent recognition unit 32 can be specifically used to use the intent recognition model and a preset prompt word library to perform intent recognition on the dialogue information input by the user and obtain the intent recognition result.

[0046] In some embodiments, the intent recognition unit 32 may also be specifically used to perform preliminary intent recognition on the dialogue information input by the user using the intent recognition big model, and obtain the confidence level of the dialogue information having different intents; if the confidence level does not meet the preset requirements, it is determined that the intent of the dialogue information is unclear, and the dialogue information is matched with prompt words using a preset prompt word library, and the intent recognition result is determined according to the prompt word matching result.

[0047] In some embodiments, the target business service intelligence body embeds a large business processing model, and the business processing unit 33 includes an extraction module and an interaction module.

[0048] The extraction module can be used to identify and extract key business parameters from the user-input dialogue information using the business processing big model.

[0049] The interaction module can be used to interact with the business system in real time based on the key business parameters in order to obtain the business data processing results.

[0050] In some embodiments, the interaction module may be specifically used to determine the communication methods supported by the business system; construct a business processing request containing the key business parameters based on the key business parameters and the interface document provided by the business system, and send the business processing request to the business system based on the communication method; and receive the business data processing results fed back by the business system based on the business processing request.

[0051] In some embodiments, the business processing unit 33 may be specifically used to enable the data interaction node of the business workflow where the target business service agent is located to directly interact with the business system in real time in order to obtain the business data processing result.

[0052] In some embodiments, the business processing unit 33 may also be specifically used to identify placeholders in a preset dialogue template; query a preset placeholder-to-business data key name mapping table based on the placeholders to determine the business data key name corresponding to the placeholders; replace the value corresponding to the business data key name in the business data processing result with the preset dialogue template to generate a business processing response script to reply to the user.

[0053] It should be noted that other corresponding descriptions of the functional units involved in the business processing device based on a large model and multi-agent collaboration provided in this embodiment can be found in the corresponding descriptions in Figure 1, and will not be repeated here.

[0054] Based on the method shown in Figure 1, this embodiment also provides a storage medium storing a computer program, which, when executed by a processor, implements the business processing method based on a large model and multi-agent collaboration as shown in Figure 1.

[0055] Based on this understanding, the technical solution of this application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as CD-ROM, USB flash drive, mobile hard drive, etc.) and includes several instructions to cause an electronic device (such as personal computer, server, or network device, etc.) to execute the methods of various implementation scenarios of this application.

[0056] Based on the method shown in Figure 1 and the virtual device embodiment shown in Figure 4, in order to achieve the above objectives, this application embodiment also provides an electronic device, which may be a personal computer, tablet computer, server, or other network device, etc. The device includes a storage medium and a processor; the storage medium is used to store computer programs; the processor is used to execute the computer programs to implement the business processing method based on large model and multi-agent collaboration shown in Figure 1.

[0057] Optionally, the aforementioned physical devices may also include a user interface, a network interface, a camera, radio frequency (RF) circuitry, sensors, audio circuitry, a Wi-Fi module, etc. The user interface may include a display screen, input units such as a keyboard, etc., and optional user interfaces may also include USB interfaces, card reader interfaces, etc. The network interface may optionally include standard wired interfaces, wireless interfaces (such as Wi-Fi interfaces), etc.

[0058] Those skilled in the art will understand that the physical device structure provided in this embodiment does not constitute a limitation on the physical device, and may include more or fewer components, or combine certain components, or have different component arrangements.

[0059] The storage medium may also include an operating system and a network communication module. The operating system is a program that manages the hardware and software resources of the aforementioned physical device, supporting the operation of information processing programs and other software and / or programs. The network communication module is used to enable communication between the various components within the storage medium, as well as communication with other hardware and software in the information processing physical device.

[0060] Through the above description of the embodiments, those skilled in the art can clearly understand that this application can be implemented by means of software plus necessary general-purpose hardware platform, or it can be implemented by hardware.

[0061] This invention, through an intent-routing intelligent agent, identifies the user's dialogue intent and accurately routes user requests to the appropriate business service intelligent agent. Simultaneously, this invention utilizes the business service intelligent agent to invoke the corresponding business workflow and interact with the business system in real time, obtaining dynamic business data processing results to facilitate business processing. Furthermore, after obtaining the business data processing results, this invention uses a standard dialogue template to construct a response, avoiding the large model illusion problem. Moreover, this invention meets the financial industry's demand for highly controllable and reliable dialogue content, while also enabling the application of large models during business processing dialogues, fully leveraging the flexible understanding capabilities of large models to enhance the user's business processing dialogue experience.

[0062] Those skilled in the art will understand that the accompanying drawings are merely schematic diagrams of a preferred embodiment, and the modules or processes shown in the drawings are not necessarily essential for implementing this application. Those skilled in the art will understand that the modules in the apparatus of the embodiment can be distributed within the apparatus of the embodiment as described, or can be modified to be located in one or more apparatuses different from this embodiment. The modules of the above-described embodiment can be combined into one module, or further divided into multiple sub-modules.

[0063] The serial numbers in this application are for descriptive purposes only and do not represent the superiority or inferiority of any particular implementation scenario. The above disclosures are merely a few specific implementation scenarios of this application; however, this application is not limited thereto, and any variations conceived by those skilled in the art should fall within the protection scope of this application.

Claims

1. A business processing method based on a large model and multi-agent collaboration, characterized in that, include: Obtain the dialogue information entered by the user during the business processing; An intent routing agent is used to identify the intent of the user's input dialogue information, and an intent identification result is obtained. A target business service agent matching the intent identification result is then determined. Each target business service agent integrates at least one business workflow, and different business service agents integrate business workflows applicable to different business scenarios. The target business service agent invokes the corresponding business workflow to interact with the business system in real time, obtaining business data processing results. These results are then used to populate variables corresponding to a preset dialogue template, generating a business processing response script to reply to the user.

2. The method according to claim 1, characterized in that, The intent routing intelligence is embedded with a large intent recognition model. The process of using the intent routing intelligence to perform intent recognition on the dialogue information input by the user and obtain intent recognition results includes: using the large intent recognition model and a preset prompt word library to perform intent recognition on the dialogue information input by the user and obtain intent recognition results.

3. The method according to claim 2, characterized in that, The step of using the intent recognition model and a preset prompt word library to perform intent recognition on the user-input dialogue information and obtain intent recognition results includes: using the intent recognition model to perform preliminary intent recognition on the user-input dialogue information to obtain confidence levels that the dialogue information has different intents; if the confidence level does not meet preset requirements, it is determined that the intent of the dialogue information is unclear; using the preset prompt word library to perform prompt word matching on the dialogue information; and determining the intent recognition result based on the prompt word matching results.

4. The method according to claim 1, characterized in that, The target business service intelligence body embeds a large business processing model. The step of using the target business service intelligence body to call the corresponding business workflow and interact with the business system in real time to obtain business data processing results includes: using the large business processing model to identify and extract key business parameters from the dialogue information input by the user; and based on the key business parameters, performing real-time data interaction with the business system to obtain the business data processing results.

5. The method according to claim 4, characterized in that, The step of interacting with the business system in real time based on the key business parameters to obtain the business data processing results includes: determining the communication methods supported by the business system; constructing a business processing request containing the key business parameters according to the key business parameters and the interface document provided by the business system, and sending the business processing request to the business system based on the communication methods; and receiving the business data processing results fed back by the business system based on the business processing request.

6. The method according to claim 1, characterized in that, The step of using the target business service agent to call the corresponding business workflow and interact with the business system in real time to obtain the business data processing result includes: enabling the data interaction node of the business workflow where the target business service agent is located to directly interact with the business system in real time to obtain the business data processing result.

7. The method according to any one of claims 4 or 6, characterized in that, The step of using the business data processing result to fill the variables corresponding to the preset dialogue template and generate a business processing response script to reply to the user includes: identifying placeholders in the preset dialogue template; querying the preset placeholder and business data key name mapping table according to the placeholders to determine the business data key name corresponding to the placeholders; replacing the value corresponding to the business data key name in the business data processing result into the preset dialogue template to generate a business processing response script to reply to the user.

8. A business processing device based on a large model and multi-agent collaboration, characterized in that, include: The acquisition unit is used to acquire dialogue information entered by the user during the business processing. The intent recognition unit is used to perform intent recognition on the dialogue information input by the user using the intent routing agent, obtain the intent recognition result, and determine the target business service agent that matches the intent recognition result. The target business service agent integrates at least one business workflow, and different business service agents integrate business workflows suitable for different business scenarios. The business processing unit is used to utilize the target business service agent to call the corresponding business workflow and interact with the business system in real time to obtain the business data processing result, and use the business data processing result to fill the variables corresponding to the preset dialogue template to generate a business processing response script to reply to the user.

9. A storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 7.

10. An electronic device comprising a storage medium, a processor, and a computer program stored on the storage medium and executable on the processor, characterized in that, When the processor executes the computer program, it implements the method of any one of claims 1 to 7.

Images