Automatic hosting service implementation method and electronic equipment

Abstract

The embodiment of the invention discloses an automatic hosting service implementation method and electronic equipment, and the method comprises the steps: obtaining context information of a negotiation stage initiated by two negotiation parties and a target scene to which the negotiation stage belongs in response to a triggering message of a hosting intervention process; on the basis of the context information and scene knowledge information associated with the target scene, a prompt text is constructed, and the scene knowledge information comprises guiding information of solving paths corresponding to multiple problems involved in the target scene; inputting the prompt text into an artificial intelligence (AI) model to obtain a processing decision generated by the AI model based on the prompt text, so as to start the trusteeship intervention process based on the processing decision; wherein the processing decision generated by the AI model based on the prompt text is a processing decision generated by understanding the user intention of the negotiation initiating user based on context information in the prompt text and performing reasoning based on guiding information in the prompt text. According to the embodiment of the invention, more flexible automatic intervention processing can be realized.

Description

Technical Field

[0001] The present application relates to the field of information processing technology, and in particular to an implementation method and electronic device for automated hosting services. Background Art

[0002] As a mechanism for handling issues between consumers and merchants, the PBC (Platform-Merchant-Consumer) platform introduces the role of the platform as an intermediary to follow up on issues where consumers and merchants cannot reach an agreement on disagreements. The platform follow-up process involves relevant business processes and rules. Therefore, initially the platform usually completes the follow-up work manually. However, since it requires a lot of manpower and time costs, it can be considered to use automation to replace manual work to improve efficiency and reduce costs.

[0003] In order to realize the above-mentioned automatic processing, in the prior art, the platform intervention process is intercepted in the PBC, and scene recognition is performed based on hard coding or some configurable rules during the interception process, and it is determined whether to directly end the automatic processing or transfer it to manual follow-up according to the recognition result.

[0004] The above method can achieve automation to a certain extent, but its disadvantages include at least: when there are new scenarios or rules, they need to be integrated through hard coding or feature development, which is costly. In addition, completely strict rule judgment may make the processing method too rigid and may not even achieve the mediation effect. Summary of the invention

[0005] The present application provides an implementation method and electronic device for automated hosting services, which can achieve more flexible automated intervention processing.

[0006] This application provides the following solutions:

[0007] A method for implementing an automated hosting service, comprising:

[0008] In response to a trigger message of the managed intervention process, obtaining context information of the negotiation phase initiated by the negotiating parties and a target scenario to which the negotiation phase belongs;

[0009] Based on the context information and the scene knowledge information associated with the target scene, construct a prompt text, wherein the scene knowledge information includes guiding information of solution paths corresponding to various problems involved in the target scene;

[0010] Inputting the prompt text into an artificial intelligence (AI) model to obtain a processing decision generated by the AI ​​model based on the prompt text, so as to start the managed intervention process based on the processing decision;

[0011] Among them, the processing decision generated by the AI ​​model based on the prompt text is based on understanding the user intention of the negotiation initiating user based on the contextual information in the prompt text, and the processing decision is generated after reasoning based on the guiding information in the prompt text.

[0012] The scenario knowledge information also includes: roles and responsibilities information required in the target scenario.

[0013] Among them, the responsibility information includes identifying the emotions of the buyer user, so that the AI ​​model also identifies the emotions of the buyer user in the process of understanding the context information to identify the buyer user's intentions, and makes inference decisions based on the identified emotions and the guiding information.

[0014] The guiding information includes: information about available tools, as well as function introductions and applicable conditions of the tools, so that the AI ​​model can determine whether a tool needs to be used to obtain information based on the prompt text, and if so, output suggestion information on using the tool;

[0015] The method further comprises:

[0016] After receiving the recommendation information for using the tool output by the AI ​​model, a call is initiated to the tool so that after obtaining relevant information by calling the tool, it is provided to the AI ​​model for further reasoning and decision-making.

[0017] Among them, after obtaining relevant information by calling the tool, a prompt text for communicating with the AI ​​model is reconstructed based on the context information, the scene knowledge information and the relevant information, so that the AI ​​model can re-make inference decisions based on the new prompt text.

[0018] After determining the target scenario to which the negotiation stage belongs, the following steps are also included:

[0019] Determine whether the target scenario is associated with a pre-configured strong rule. If so, divert the process to the process of automated intervention processing through the strong rule. Otherwise, trigger the step of using the AI ​​model to assist in completing the automated intervention processing process.

[0020] Among them, it also includes:

[0021] If the inference result returned by the AI ​​model is that the problem cannot be solved, a process of manual intervention is triggered.

[0022] Among them, it also includes:

[0023] The data in the processing flow of the AI ​​model's reasoning and decision-making is stored so as to filter out high-quality data from the stored data as training samples and fine-tune the AI ​​model.

[0024] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of any of the methods described above.

[0025] An electronic device, comprising:

[0026] one or more processors; and

[0027] A memory associated with the one or more processors, the memory being used to store program instructions, wherein the program instructions, when read and executed by the one or more processors, execute the steps of any of the methods described above.

[0028] A computer program product comprises a computer program / computer executable instructions, wherein the computer program / computer executable instructions implement the steps of any of the aforementioned methods when executed by a processor in an electronic device.

[0029] According to the specific embodiments provided in this application, this application discloses the following technical effects:

[0030] Through the embodiment of the present application, in the process of the platform party intervening in the negotiation stage between the buyer and the seller, the AI ​​model can be used to assist in completing the automated intervention process to reduce the reliance on the platform party's manual customer service. Among them, in the specific implementation process, the scene knowledge information can be provided in advance for the specific scene, which may include the guiding information of the solution path corresponding to the various problems involved in the target scene. In this way, in the process of intervening specifically in the negotiation stage between the buyer and the seller, the main program can obtain the context information related to the negotiation stage and determine the target scene to which the negotiation stage belongs. After that, a prompt text (prompt) for dialogue with the AI ​​model can be constructed based on the context information and the scene script corresponding to the target scene. Correspondingly, the AI ​​model can generate a processing decision based on the prompt text so as to start the managed intervention process based on the processing decision. Among them, the processing decision generated by the AI ​​model based on the prompt text can specifically be to understand the user intention of the negotiation initiator based on the context information in the prompt text, and to generate a processing decision based on the guiding information in the prompt text after reasoning. In this way, the AI ​​model's ability to understand user intent can be utilized to achieve more flexible automated intervention processing. In addition, the usability of the solutions generated by the AI ​​model can also be improved by setting guiding information on the solution path in the scenario knowledge information.

[0031] In a preferred implementation, the guiding information in the scene knowledge information may include information about available tools, as well as the functional introduction and applicable condition information of the tools. In this way, the AI ​​model can determine whether a tool needs to be used to obtain more relevant information. If necessary, the main program calls the tool, obtains the information, and then provides it to the AI ​​model for further reasoning and decision-making. In this way, the decoupling between the tool call and the judgment logic can be achieved, and the judgment process can also make full use of the advantages of the AI ​​model in understanding user intent and other aspects.

[0032] Of course, any product implementing the present application does not necessarily need to achieve all of the advantages described above at the same time. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0034] Figure 1 It is a schematic diagram of the system architecture provided by the embodiment of the present application;

[0035] Figure 2 is a flow chart of the method provided in the embodiment of the present application;

[0036] Figure 3 is a schematic diagram of a device provided in an embodiment of the present application;

[0037] Figure 4 It is a schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0038] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. All other embodiments obtained by ordinary technicians in this field based on the embodiments in the present application belong to the scope of protection of this application.

[0039] First of all, it should be noted that the reason why the processing results are too rigid and even fail to achieve the effect of mediation when implementing automatic hosting through preset strong rules is that in the process of mediating the negotiation between buyers and sellers, the intentions and emotions of consumers are very important. However, when providing solutions through strong rule matching, no meaningful results are identified from the perspectives of consumers' intentions and emotions. Therefore, the solutions provided may be difficult to play a role in mediation, and may even cause disgust among consumers. For example, if the consumer user is very emotional, a special processing method may be required, or manual intervention may be required. However, under the strong rule method, the consumer's emotions cannot be perceived, so it may be difficult to provide a solution that satisfies the consumer.

[0040] Therefore, in the embodiments of the present application, the AI ​​(Artificial Intelligence) Large Language Model (LLM) can be used to assist in the completion of the automated platform intervention process in terms of intent understanding and reasoning. Among them, LLM is a language understanding and generation model built on deep learning technology. It can generate coherent text with semantic and grammatical correctness through training of large-scale text data, and has certain reasoning capabilities, and can perform a wide range of tasks, such as text summarization, translation, sentiment analysis, etc. The embodiments of the present application can assist in the completion of the automated platform intervention process with the help of the above-mentioned capabilities of the LLM model. Of course, the LLM model is a general term for a class of models. In the specific implementation, other models besides the LLM model can also be used to implement it, as long as they have the capabilities of intent understanding, text generation, etc., which are collectively referred to as AI models in the embodiments of the present application.

[0041] Of course, if the pre-trained AI model is directly applied to the platform intervention scenario in the embodiment of the present application, there may be a situation where the AI ​​model cannot accurately provide a solution. In addition, in the process of applying the AI ​​model, it also involves the construction of prompts (prompt text, used to provide the AI ​​model with input information as input information of the AI ​​model, guide the content generation of AI models such as LLM, and hope that the AI ​​model will feedback the expected results). In general, it may be necessary to manually input a specific prompt. However, in the scenario of the embodiment of the present application, it is obviously impossible to complete the input of the prompt manually. Therefore, how to automatically construct a prompt and guide the AI ​​model to make relatively accurate decisions is a problem that needs to be considered.

[0042] Based on the above problems, the embodiment of the present application provides a corresponding solution, in which scenario knowledge information can be prepared in advance for specific scenarios. The scenarios for negotiation between the buyer and the seller may include refunds, logistics and other different scenarios. Different scenario knowledge information can be prepared for different scenarios, or scenario knowledge information can be prepared for scenarios with high incidence such as refunds, etc. In addition, some basic knowledge information can be prepared. When some new scenarios and other situations arise, the basic knowledge information can be used to promote the process.

[0043] Among them, the specific scenario knowledge information may include guiding information on the solution paths corresponding to the various problems involved in the target scenario. Of course, in an optional manner, the scenario knowledge information may also include the roles and their responsibilities required in the target scenario. For example, the specific role may usually be a "platform customer service staff", etc., whose responsibilities may be "locating members' problems and demands to provide corresponding SOP solutions. In the process, you can also view the views of member-related information to assist in locating members' problems, and track the progress of member problem handling after providing solutions to finally complete member services." Guiding information on the solution path can be based on some knowledge information (for example, what operations (including querying some status information, etc.) the platform has performed for various problems in the process of manual intervention, what solutions have been given, etc.) to provide guidance for the AI ​​model, including what solutions can be adopted in various possible situations or states in the scenario, or what more information can be queried first, and then what solutions can be given based on the queried information, etc.

[0044] In addition, the contextual information of the negotiation stage between the buyer and the seller (including associated order information, logistics information, conversation content information, etc.) can also be obtained. In this way, based on the scenario knowledge information and the above contextual information, the main program of the automated intervention system can automatically construct a prompt for communicating with the AI ​​model. Afterwards, the AI ​​model can make inference decisions based on the specific prompt.

[0045] In the process of reasoning and decision-making, the AI ​​model can also determine whether it needs to obtain additional information, such as querying logistics status information, etc. If necessary, it can also determine the tools that can be called. At this time, the content returned by the AI ​​model can be the suggestion information about calling a certain tool. After that, the main program can call the corresponding tool, and after obtaining specific information, it can be provided to the AI ​​model for further reasoning and decision-making.

[0046] Of course, in the specific implementation, for some scenarios, the platform can also automatically intervene and process through strong rules. Therefore, after the platform is triggered to intervene in the negotiation stage between the buyer and the seller, it can also be diverted according to the scenario. Some scenarios can be automatically processed through strong rules, and some other scenarios can be automatically processed with the assistance of AI models. Among them, which scenarios are suitable for strong rules can be pre-configured.

[0047] From the perspective of system architecture, see Figure 1 , the embodiment of the present application can provide an automated platform intervention processing service in the PBC system. From a functional perspective, it can be divided into three domains: the business domain, which is mainly responsible for the initiation of PBC, the determination of reflux rules, the triggering of automated hosting and the control of hosting status, etc.; the work order domain, which is mainly responsible for the creation, driving, completion of work orders, and the triggering of automated tasks (the source of triggering hosting); and the intelligent domain, which is mainly responsible for the control of AI models, agent execution, and the execution of related scenario hosting process tools, etc. In specific implementation, PBC can be initiated by the business domain, that is, the platform party can be triggered to intervene in the process according to the negotiation situation between the buyer and the seller and the corresponding business rules, and the corresponding information reflux can be supported. After the platform party intervenes in the process, the work order domain can create a work order and drive the work order, and the automated hosting trigger can be performed through event notification. After that, the business domain can construct a prompt and call the AI ​​model. LLM-Agent (LLM agent) controls the response and reply of the agent through the prompt project. Roles, instructions, permissions, responses and replies to contexts, etc. can be designed in the prompt. Sufficient prompt + knowledge can enable Agent to go from a conversational chatbot to a goal-driven workflow and task automation. The process engine is an underlying support platform that provides process processing development. The correctness of the process engine can be verified through upper-level applications.

[0048] The specific implementation scheme provided in the embodiments of the present application is described in detail below.

[0049] First, the present application embodiment provides a method for implementing an automated hosting service. Figure 2 , the method may include:

[0050] S201: In response to a trigger message of a managed intervention process, context information of a negotiation phase initiated by two negotiating parties and a target scenario to which the negotiation phase belongs are obtained.

[0051] In specific implementation, when the buyer needs to refund or return goods, or finds logistics problems, he or she can usually first negotiate with the seller to resolve the situation. If both parties can reach an agreement, refunds or returns can be made in the usual way, or the seller can make compensation, etc. However, if the negotiation between the buyer and the seller fails, the platform can intervene. In specific implementation, the event of the negotiation failure between the buyer and the seller can trigger an automated intervention task work order and enter the platform intervention process.

[0052] After the platform intervenes, the main program of the automated intervention process can obtain the context information related to the negotiation stage, including the order information involved in the specific negotiation stage (including the goods associated with the order, etc.), logistics information (mainly logistics order numbers, etc.), the content of the conversation between the two parties, etc. In addition, the target scenario associated with the negotiation stage can also be determined based on this context information, for example, whether it is a refund, return or exchange scenario, or a logistics scenario, etc.

[0053] After determining the target scene, in an optional manner, it is also possible to determine whether the target scene is associated with a pre-configured strong rule. If so, it can be diverted to the process of automatic intervention processing through the strong rule. Otherwise, it can trigger the step of using the AI ​​model to assist in completing the automatic intervention processing process. Of course, in another manner, all scenes can also be assisted by the AI ​​model to complete the automatic processing, which is not limited here.

[0054] S202: Constructing a prompt text based on the context information and the scenario knowledge information associated with the target scenario, wherein the scenario knowledge information includes guiding information on solution paths corresponding to various problems involved in the target scenario.

[0055] After determining the target scenario and determining that it needs to be assisted by the AI ​​model, the scenario knowledge information associated with the target scenario can also be determined, wherein the scenario knowledge information can be pre-configured according to the historical intervention records of the specific scenario or the experience of some operators, etc., which may include guiding information on the solution paths corresponding to the various problems involved in the target scenario. Alternatively, the scenario knowledge information may also include the required roles (mainly the roles that the AI ​​model needs to play) and their responsibilities information. For example, the specific role may usually be the customer service staff of the platform, etc., and the specific responsibility is to mediate disputes between buyers and sellers, and provide solutions that are acceptable to both parties. In addition, in the preferred implementation method, the specific responsibility information may also include the identification of the buyer's emotions, so as to guide the AI ​​model to identify the emotions of the negotiation initiating user in the process of understanding the context information to identify the intention of the negotiation initiating user, so as to make reasoning decisions based on the identified emotions.

[0056] In addition, in order to help the AI ​​model make inference decisions, it is also possible to provide guidance information on the solution paths corresponding to various problems involved in the target scenario in the scenario knowledge information. For example, specific guidance information may include whether more information is needed (for example, including querying order status, logistics status, etc.), what solutions can be adopted when it is recognized that the user has a certain intention, and so on.

[0057] That is to say, in an embodiment of the present application, the AI ​​model can determine whether it is necessary to obtain more information. If necessary, the information can be obtained first and then a decision can be made. In specific implementation, in order to facilitate information acquisition, a tool for obtaining specific information can be provided in advance. This tool can exist in the form of a functional module or service. When it is determined that a certain aspect of information needs to be obtained, the information can be obtained by calling the corresponding tool. In this case, the scene knowledge information can also include information about the available tools, as well as information such as the functional introduction and applicable conditions corresponding to the tool, so that the AI ​​model can determine whether a tool needs to be used for information acquisition based on the prompt text. If so, it outputs recommendation information about the use of the tool.

[0058] After determining the context information and scene knowledge information, a prompt for communicating with the AI ​​model can be constructed based on this information. Specifically, the scene knowledge information may include some placeholders, such as order information, conversation content between buyers and sellers, etc., which can exist in the scene knowledge information as placeholders. After obtaining the specific context information, the scene knowledge information can be filled in and then a specific prompt can be generated.

[0059] For example, in a certain application example, the prompt content may include:

[0060] "'You are a customer service clerk (role) on the platform. You mainly provide services to members. You provide corresponding SOP solutions by locating members' problems and demands. During the process, you can also view the views of member-related information to assist in locating members' problems. After providing solutions, you track the progress of member problem handling and finally complete member services (responsibilities).'

[0061] You can use the following tools to assist the service process. You can use one or more tools at the same time:

[0062] View tool (the tool name or calling address and other information may be given later) is used to view members' related orders, refunds, logistics, work orders and other information to help locate problems;

[0063] Other tools...

[0064] Output content format requirements

[0065] Current context input parameter: {context content}...'"

[0066] Of course, the above prompt is only used for illustration and should not be regarded as limiting the scope of protection of this application.

[0067] S203: Input the prompt text into an artificial intelligence AI model to obtain a processing decision generated by the AI ​​model based on the prompt text, so as to initiate the hosted intervention process based on the processing decision; wherein the processing decision generated by the AI ​​model based on the prompt text is a processing decision generated after understanding the user intention of the negotiation initiating user based on the contextual information in the prompt text and reasoning based on the guiding information in the prompt text.

[0068] By outputting the above prompt to the AI ​​large language model, the AI ​​model can be guided to make corresponding reasoning decisions and provide corresponding solutions. Accordingly, after receiving the prompt, the AI ​​model can understand the context information of the negotiation stage according to the information in the prompt to identify the buyer's user intention. On this basis, it can make reasoning decisions based on the guiding information and complete its corresponding responsibilities to assist in completing the automated platform intervention process.

[0069] Among them, since the guiding information in the scene knowledge information can include information about available tools, as well as the functional introduction and applicable condition information corresponding to the tools, the AI ​​model can also judge whether it is necessary to use a certain tool to obtain information based on the prompt text. If so, it can output suggestion information about the use of the tool. That is, the AI ​​model decides whether it is necessary to call a certain tool, and if so, it can output suggestion information about calling the tool. Correspondingly, after receiving the suggestion information for using the tool output by the AI ​​model, the main program can initiate a call to the tool, and after obtaining relevant information by calling the tool, it can provide it to the AI ​​model for further reasoning and decision-making. In other words, the AI ​​model can decide whether it is necessary to call a certain tool. If necessary, the main program executes the specific calling process, and the query results and other information obtained can be further provided to the AI ​​large language model for further reasoning and decision-making.

[0070] It should be noted here that, in the case of a specific tool for obtaining relevant information, one implementation method may be that after the main program obtains the context information in the negotiation stage between the buyer and the seller, the main program determines whether it is necessary to obtain more information based on the context information, for example, whether it is necessary to query the logistics status, etc. If necessary, after calling the specific tool to obtain the corresponding information, as a supplement to the context information, the prompt is constructed together with the scene script, and then the AI ​​model makes an inference decision. However, this method means that it is necessary to write code in the main program to complete the above judgment and the calling of the tool. In the embodiment of the present application, an implementation method is adopted in which the AI ​​model judges whether a tool needs to be called. The main program only needs to call the tool according to the suggestions given by the AI ​​large language model, so that the calling of the tool is decoupled from the judgment of whether the tool needs to be called, which is conducive to reducing the difficulty of implementation. In addition, when the AI ​​model is used to judge whether a tool needs to be called, the ability of the AI ​​model in understanding user intent and other aspects can also be utilized in the judgment process, and the accuracy can also be improved compared to the judgment made by the hard coding method in the main program.

[0071] It should also be noted that in an embodiment of the present application, after the main program obtains relevant information by calling the tool, it can re-construct a prompt text for dialogue with the AI ​​model based on the context information of the negotiation stage between the buyer and the seller, the scene knowledge information and the relevant information, so that the AI ​​model can re-infer and decide based on the new prompt text. In other words, the interaction between the main program and the AI ​​model can adopt a single-round interaction method. For the AI ​​model, even if it is determined that a tool needs to be called to obtain more information, the suggestion information of the calling tool can be used as a conclusion to end the dialogue process. After obtaining more information, the main program can be reassembled into a prompt together with other context information, scene knowledge information, etc., and re-initiate interaction with the AI ​​model. After re-initiating a dialogue with the AI ​​model, the AI ​​model re-infers and makes a decision. If it is no longer necessary to obtain other information, the conclusion of the reasoning decision can be given. For example, the negotiation situation between the buyer and the seller can be summarized, and conclusive content such as solutions can be given. Of course, in the specific implementation, a conversation can also be created between the main program and the AI ​​model, so that the two can interact through multiple rounds of conversations. At this time, after the AI ​​model gives a suggestion to call a tool, the main program calls the tool to obtain relevant information. After that, the relevant information can be directly input into the AI ​​model, and the AI ​​model can automatically combine the historical conversation content for further reasoning and decision-making.

[0072] Of course, in actual applications, there are also situations where the AI ​​model cannot provide a solution, or it is found that the buyer's emotions are very excited and manual intervention is needed, etc. At this time, the AI ​​model can return conclusive content about the failure to solve the problem. Correspondingly, if the reasoning result returned by the AI ​​model is that the problem cannot be solved, it can trigger the process of manual intervention, generate a work order for manual intervention, and assign the work order task to manual customer service, etc.

[0073] Among them, in the embodiment of the present application, in the process of platform intervention assisted by the AI ​​model, although the AI ​​model is required to assume that it is the customer service staff of the platform when constructing the prompt, the AI ​​model is not required to communicate with the buyer and seller users. It only needs to make reasoning decisions based on the context of the negotiation stage between the buyer and seller and provide a solution. If a solution cannot be provided, it returns the conclusion that no solution has been provided.

[0074] It should be noted that in the specific implementation, for the above-mentioned cases where the AI ​​model performs reasoning and decision-making, snapshots of the data generated in all intermediate stages can be stored. Afterwards, senior personnel can evaluate this part of the stored processing flow data to generate high-quality training samples, and then fine-tune the large model (Supervised Fine-Tun-ing, SFT), thereby improving the model's reasoning and decision-making capabilities in some scenarios.

[0075] In summary, through the embodiments of the present application, in the process of the platform intervening in the negotiation stage between the buyer and the seller, the AI ​​model can be used to assist in completing the automated intervention process to reduce the reliance on the platform's manual customer service. Among them, in the specific implementation process, scene knowledge information can be provided in advance for specific scenarios, which may include guiding information on the solution paths corresponding to the various problems involved in the target scenario. In this way, in the process of intervening specifically in the negotiation stage between the buyer and the seller, the main program can obtain context information related to the negotiation stage and determine the target scene to which the negotiation stage belongs. After that, a prompt text (prompt) for dialogue with the AI ​​model can be constructed based on the context information and the scene script corresponding to the target scene. Correspondingly, the AI ​​model can generate a processing decision based on the prompt text so as to start the managed intervention process based on the processing decision. Among them, the processing decision generated by the AI ​​model based on the prompt text can specifically be to understand the user intention of the negotiation initiator based on the context information in the prompt text, and to generate a processing decision based on the guiding information in the prompt text after reasoning. In this way, the AI ​​model's ability to understand user intent can be utilized to achieve more flexible automated intervention processing. In addition, the usability of the solutions generated by the AI ​​model can also be improved by setting guiding information on the solution path in the scenario knowledge information.

[0076] In a preferred implementation, the guiding information in the scene knowledge information may include information about available tools, as well as the functional introduction and applicable condition information of the tools. In this way, the AI ​​model can determine whether a tool needs to be used to obtain more relevant information. If necessary, the main program calls the tool, obtains the information, and then provides it to the AI ​​model for further reasoning and decision-making. In this way, the decoupling between the tool call and the judgment logic can be achieved, and the judgment process can also make full use of the advantages of the AI ​​model in understanding user intent and other aspects.

[0077] It should be noted that the embodiments of the present application may involve the use of user data. In actual applications, user-specific personal data can be used in the scheme described herein within the scope permitted by applicable laws and regulations, subject to the requirements of applicable laws and regulations of the country where the user is located (for example, with the user's explicit consent, effective notification to the user, etc.).

[0078] Corresponding to the above method embodiment, the present application embodiment also provides an implementation device of an automated hosting service, see Figure 3 , the device may include:

[0079] The context and scenario information acquisition unit 301 is used to obtain the context information of the negotiation phase initiated by the negotiating parties and the target scenario to which the negotiation phase belongs in response to the trigger message of the managed intervention process;

[0080] A prompt text construction unit 302 is used to construct a prompt text based on the context information and the scene knowledge information associated with the target scene, wherein the scene knowledge information includes guiding information of solution paths corresponding to various problems involved in the target scene;

[0081] A prompt text input unit 303, used to input the prompt text into an artificial intelligence AI model to obtain a processing decision generated by the AI ​​model based on the prompt text, so as to start the managed intervention process based on the processing decision;

[0082] Among them, the processing decision generated by the AI ​​model based on the prompt text is based on understanding the user intention of the negotiation initiating user based on the contextual information in the prompt text, and the processing decision is generated after reasoning based on the guiding information in the prompt text.

[0083] In a specific implementation, the scenario knowledge information may further include: roles and responsibilities information required in the target scenario.

[0084] Specifically, the responsibility information includes identifying the buyer user's emotions, so that the AI ​​model also identifies the buyer user's emotions in the process of understanding the context information to identify the buyer user's intentions, and makes inference decisions based on the identified emotions and the guiding information.

[0085] The guiding information may specifically include: information about available tools, as well as a functional introduction and applicable condition information corresponding to the tools, so that the AI ​​model can determine whether a tool needs to be used to obtain information based on the prompt text, and if so, output suggestion information on using the tool;

[0086] At this time, the device may further include:

[0087] The tool calling unit is used to initiate a call to the tool after receiving the recommendation information for using the tool output by the AI ​​model, so as to provide the relevant information obtained by calling the tool to the AI ​​model for further reasoning and decision-making.

[0088] In addition, the device may also include:

[0089] The prompt text reconstruction unit is used to reconstruct the prompt text for communicating with the AI ​​model based on the context information, the scene knowledge information and the relevant information after obtaining the relevant information by calling the tool, so that the AI ​​model can re-make reasoning decisions based on the new prompt text.

[0090] Furthermore, the device may further include:

[0091] A judgment unit is used to determine whether the target scene to which the negotiation stage belongs is associated with a pre-configured strong rule after determining the target scene to which the negotiation stage belongs. If so, the process is diverted to the process of automatic intervention processing through the strong rule; otherwise, it triggers the step of using the AI ​​model to assist in completing the automatic intervention processing process.

[0092] The manual intervention process triggering unit is used to trigger the manual intervention process if the reasoning result returned by the AI ​​model is that the problem cannot be solved.

[0093] A data storage unit is used to store data in the processing flow of the AI ​​model for reasoning and decision-making, so as to screen out high-quality data from the stored data as training samples and fine-tune the AI ​​model.

[0094] In addition, an embodiment of the present application further provides a computer-readable storage medium on which a computer program is stored, and when the program is executed by a processor, the steps of any one of the methods in the aforementioned method embodiments are implemented.

[0095] And an electronic device, comprising:

[0096] one or more processors; and

[0097] A memory associated with the one or more processors, the memory being used to store program instructions, wherein the program instructions, when read and executed by the one or more processors, execute the steps of the method described in any one of the aforementioned method embodiments.

[0098] A computer program product includes a computer program / computer executable instructions, which, when executed by a processor in an electronic device, implement the steps of the method described in the aforementioned method embodiment.

[0099] in, Figure 4The architecture of the electronic device is shown as an example, which may include a processor 410, a video display adapter 411, a disk drive 412, an input / output interface 413, a network interface 414, and a memory 420. The processor 410, the video display adapter 411, the disk drive 412, the input / output interface 413, the network interface 414, and the memory 420 may be communicatively connected via a communication bus 430.

[0100] Among them, the processor 410 can be implemented by a general-purpose CPU (Central Processing Unit, processor), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, etc., to execute relevant programs to implement the technical solution provided in this application.

[0101] The memory 420 can be implemented in the form of ROM (Read Only Memory), RAM (Random Access Memory), static storage device, dynamic storage device, etc. The memory 420 can store an operating system 421 for controlling the operation of the electronic device 400, and a basic input and output system (BIOS) for controlling the low-level operation of the electronic device 400. In addition, a web browser 423, a data storage management system 424, and a hosting service processing system 425, etc. can also be stored. The above-mentioned hosting service processing system 425 can be an application program that specifically implements the operations of the aforementioned steps in the embodiment of the present application. In short, when the technical solution provided in the present application is implemented by software or firmware, the relevant program code is stored in the memory 420 and is called and executed by the processor 410.

[0102] The input / output interface 413 is used to connect the input / output module to realize information input and output. The input / output module can be configured in the device as a component (not shown in the figure), or it can be externally connected to the device to provide corresponding functions. The input device may include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device may include a display, a speaker, a vibrator, an indicator light, etc.

[0103] The network interface 414 is used to connect to a communication module (not shown) to realize communication interaction between the device and other devices. The communication module can realize communication through a wired mode (such as USB, network cable, etc.) or a wireless mode (such as mobile network, WIFI, Bluetooth, etc.).

[0104] The bus 430 comprises a pathway for transmitting information between the various components of the device (eg, the processor 410, the video display adapter 411, the disk drive 412, the input / output interface 413, the network interface 414, and the memory 420).

[0105] It should be noted that, although the above device only shows a processor 410, a video display adapter 411, a disk drive 412, an input / output interface 413, a network interface 414, a memory 420, a bus 430, etc., in the specific implementation process, the device may also include other components necessary for normal operation. In addition, it can be understood by those skilled in the art that the above device may also only include components necessary for implementing the solution of the present application, and does not necessarily include all the components shown in the figure.

[0106] It can be known from the description of the above implementation methods that those skilled in the art can clearly understand that the present application can be implemented by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be essentially or partly contributed to the prior art in the form of a software product, which can be stored in a storage medium such as ROM / RAM, a magnetic disk, an optical disk, etc., and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in the various embodiments of the present application or certain parts of the embodiments.

[0107] Each embodiment in this specification is described in a progressive manner, and the same or similar parts between the embodiments can refer to each other, and each embodiment focuses on the differences from other embodiments. In particular, for the system or system embodiment, since it is basically similar to the method embodiment, the description is relatively simple, and the relevant parts can refer to the partial description of the method embodiment. The system and system embodiments described above are merely schematic, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without creative work.

[0108] The above is a detailed introduction to the implementation method and electronic device of the automated hosting service provided by the present application. This article uses specific examples to illustrate the principles and implementation methods of the present application. The description of the above embodiments is only used to help understand the method and core ideas of the present application. At the same time, for those skilled in the art, according to the ideas of the present application, there will be changes in the specific implementation methods and application scope. In summary, the content of this specification should not be understood as limiting the present application.

Claims

1. A method for implementing an automated hosting service, comprising: In response to a trigger message of the managed intervention process, obtaining context information of the negotiation phase initiated by the negotiating parties and a target scenario to which the negotiation phase belongs; Based on the context information and the scene knowledge information associated with the target scene, construct a prompt text, wherein the scene knowledge information includes guiding information of solution paths corresponding to various problems involved in the target scene; Inputting the prompt text into an artificial intelligence (AI) model to obtain a processing decision generated by the AI ​​model based on the prompt text, so as to start the managed intervention process based on the processing decision; Among them, the processing decision generated by the AI ​​model based on the prompt text is based on understanding the user intention of the negotiation initiating user based on the contextual information in the prompt text, and the processing decision is generated after reasoning based on the guiding information in the prompt text.

2. The method according to claim 1, characterized in that The scenario knowledge information also includes: roles and responsibilities information required in the target scenario.

3. The method according to claim 2, characterized in that The responsibility information includes identifying the buyer user's emotions, so that the AI ​​model can understand the buyer user's intentions based on the context information, identify the buyer user's emotions, and make inference decisions based on the identified emotions and the guiding information.

4. The method according to claim 1, characterized in that: The guiding information includes: information about available tools, as well as a description of the functions and applicable conditions of the tools, so that the AI ​​model can determine whether a tool needs to be used to obtain information based on the prompt text, and if so, output suggestion information on using the tool; The method further comprises: After receiving the recommendation information for using the tool output by the AI ​​model, a call is initiated to the tool so that after obtaining relevant information by calling the tool, it is provided to the AI ​​model for further reasoning and decision-making.

5. The method according to claim 4, characterized in that After obtaining relevant information by calling the tool, a prompt text for communicating with the AI ​​model is reconstructed based on the context information, the scene knowledge information and the relevant information, so that the AI ​​model can re-make inference decisions based on the new prompt text.

6. The method according to claim 1, characterized in that After determining the target scenario to which the negotiation stage belongs, the following steps are also included: Determine whether the target scenario is associated with a pre-configured strong rule. If so, divert the process to the process of automated intervention processing through the strong rule. Otherwise, trigger the step of using the AI ​​model to assist in completing the automated intervention processing process.

7. The method according to claim 1, characterized in that Also includes: If the inference result returned by the AI ​​model is that the problem cannot be solved, a process of manual intervention is triggered.

8. The method according to any one of claims 1 to 7, characterized in that: Also includes: The data in the processing flow of the AI ​​model's reasoning and decision-making is stored so as to filter out high-quality data from the stored data as training samples and fine-tune the AI ​​model.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps of the method described in any one of claims 1 to 8 are implemented.

10. An electronic device, characterized in that: include: one or more processors; as well as A memory associated with the one or more processors, the memory being used to store program instructions, wherein the program instructions, when read and executed by the one or more processors, execute the steps of the method described in any one of claims 1 to 8.

11. A computer program product comprising a computer program / computer executable instructions, characterized in that: When the computer program / computer executable instructions are executed by a processor in an electronic device, the steps of the method according to any one of claims 1 to 8 are implemented.

Images