Session system, session method, electronic equipment and computer readable storage medium
By storing the historical session content between the AI agent and the customer in the database, the session content sharing between the agent is realized, and the problem of AI agents giving inconsistent and inaccurate solutions is solved, and the customer experience and decision-making quality is improved.
Patent Information
- Application Number
- CN202510040391.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2025-01-10
- Publication Date
- 2025-04-25
AI Technical Summary
The solutions given by existing AI agents when facing similar problems are inconsistent, inaccurate and incomplete, resulting in poor customer experience.
By storing historical session content between different agents and customers in the database, the session content sharing between agents is realized, breaking the 'knowledge silo', and generating solutions based on historical session content and current session content.
Improves the ability of agents to provide consistent solutions, enhances the utilization of knowledge, and improves customer experience and decision-making quality.
Smart Images
Figure CN119886356A_ABST
Abstract
Description
Technical Field
[0001] The present application belongs to the field of computer technology, and specifically relates to a conversation system, a conversation method, an electronic device, and a computer-readable storage medium. Background Art
[0002] AI (Artificial Intelligence) Agent refers to an intelligent system that can autonomously perform tasks or make decisions in a certain environment. The core of AI Agent is Large Language Models (LLM). Among them, large language models have the functions of understanding natural language, performing text analysis, generating text responses, and assisting decision-making.
[0003] Currently, when using AI agents to solve customer problems, the AI agents will directly use large language models to provide relevant solutions. Since the large language models of different AI agents are all private domain large language models, this will lead to inconsistent solutions given by different AI agents for similar problems, resulting in poor customer experience and incomplete and inaccurate solutions given by AI agents. Summary of the invention
[0004] In view of this, the purpose of the present application is to provide a conversation system, a conversation method, an electronic device and a computer-readable storage medium to improve the problems that the solutions given by current AI agents to similar problems are inconsistent, and the solutions given are inaccurate and incomplete.
[0005] The embodiment of the present application is implemented as follows:
[0006] In a first aspect, an embodiment of the present application provides a conversation system, comprising: a conversation management module and an agent. The conversation management module is used to query a target historical conversation content related to a first query request from a database, wherein the database stores historical conversation content between different agents and customers, and the historical conversation content includes questions raised by the customer and solutions given by the agent to the questions; the agent is used to send the first query request to the conversation management module according to the customer's current conversation content, and receive a first query result returned by the conversation management module in response to the first query request, and when the first query result indicates that the target historical conversation content has been queried, a first solution is obtained based on the target historical conversation content and the current conversation content.
[0007] In the above embodiment, by storing the historical conversation contents between different intelligent agents and customers in a database, the historical conversation contents of each intelligent agent can be integrated so that the conversation contents can be shared between intelligent agents, breaking the barriers of "knowledge islands" and improving the utilization rate of knowledge; it can also enable intelligent agents to use existing solutions to answer customer questions, which not only reduces the workload and improves work efficiency, but also allows different intelligent agents to give consistent solutions to similar problems, thereby improving the customer experience; and by sharing conversation content, intelligent agents can give more comprehensive and accurate solutions and improve the quality of decision-making.
[0008] In combination with the embodiment of the first aspect, in a possible implementation manner, the intelligent agent is also used to determine the type of problem in the current session content; when the type of problem in the current session content is a complex problem, enter the slow thinking mode; wherein, entering the slow thinking mode includes: sending the first query request to the session management module according to the customer's current session content, and receiving the first query result returned by the session management module in response to the first query request, and when the first query result represents that the target historical session content is queried, obtaining a first solution based on the target historical session content and the current session content.
[0009] In the above embodiment, different thinking modes can be selected according to the type of problem in the current conversation content. For complex problems, a slow thinking mode will be entered, allowing the intelligent agent to enter a deeper analysis and consider more factors and possibilities when providing solutions. By integrating the historical conversation content of each intelligent agent, complex problems can be grasped and understood more accurately, and more comprehensive and balanced decisions can be made, thereby improving the quality of decision-making. At the same time, by referring to historical conversation content, the intelligent agent can reduce the possibility of repeatedly creating existing knowledge, thereby saving resources, and can also enable different intelligent agents to provide consistent solutions to similar problems, thereby improving customer experience.
[0010] In combination with the first aspect embodiment, in a possible implementation, the intelligent agent is also used to enter a quick thinking mode when the type of question in the current conversation content is a simple question; wherein the quick thinking mode includes: obtaining a second solution directly based on the current conversation content.
[0011] In the above embodiment, when the question in the current conversation content is a simple question, the quick thinking mode is entered, that is, a solution is obtained directly based on the current conversation content, so that the intelligent agent can respond to customer questions quickly and improve work efficiency; at the same time, for simple questions, the quick thinking mode can reduce the consumption of computing resources and save resources.
[0012] In combination with the first aspect embodiment, in a possible implementation manner, the intelligent agent is also used to obtain a first response from the customer based on the second solution; when the first response indicates that the problem in the current conversation content is not resolved, the slow thinking mode is entered.
[0013] In the above embodiment, when the solution provided in the fast thinking mode cannot solve the problem in the current conversation content, it is immediately switched to the slow thinking mode, and an in-depth analysis is performed through the slow thinking mode to provide a higher quality solution, thereby improving the customer experience.
[0014] In combination with the first aspect embodiment, in a possible implementation manner, the intelligent agent is also used to desensitize the content of the current session after the current session ends, to obtain a first historical session content including the problem and the solution given by the intelligent agent to the problem, and to send the first historical session content to the session management module for storage.
[0015] In the above embodiment, before saving the content of this session, desensitization processing is first performed to remove sensitive information in the content of this session to protect the privacy and security of the customer, reduce the risk of malicious use of the session content, and enhance data security; at the same time, after each session ends, the current session content is saved in time to update the historical session content in the database in real time, so that the current session content can be shared by different intelligent agents more quickly, which is convenient for the intelligent agents to use as a reference when answering customer questions, improve work efficiency, and ensure that the session content in the database is the latest.
[0016] In combination with the first aspect embodiment, in a possible implementation manner, the session management module is also used to, when receiving the first historical session content, query similar historical session content related to the first historical session content from the database; when the similar historical session content is queried, merge the first historical session content and the similar historical session content to obtain merged historical session content; and save the merged historical session content to the database.
[0017] In the above embodiment, when saving the first historical conversation content, if there is similar historical conversation content related to the first historical conversation content in the database, the first historical conversation content and the similar historical conversation content are merged and then saved in the database. This can ensure the consistency of the data in the database, reduce data duplication and conflict, and provide the agent with more comprehensive and complete conversation content when querying, so that the agent can be more accurate and reliable when answering customer questions; at the same time, by merging similar historical conversation content, not only can storage space be saved, but also the work efficiency of the agent can be improved.
[0018] In combination with the first aspect embodiment, in a possible implementation, the session management module is also used to, when receiving multiple similar second query requests in a target time period, fuse the query contents in the multiple second query requests to obtain a target query request; query the target historical session content related to the target query request from the database to obtain a second query result; and send the second query results to the intelligent agents that sent the second query requests respectively.
[0019] In the above embodiment, by fusing the query contents of multiple similar query requests, information from multiple perspectives can be integrated, making the queried target historical conversation content more comprehensive, thereby improving the quality and efficiency of decision-making; at the same time, by fusing multiple similar query requests, it helps to connect scattered information points, reduce knowledge islands, and promote information sharing; in addition, it can also reduce repeated processing of similar query requests, thereby improving processing efficiency and resource utilization.
[0020] In combination with the embodiment of the first aspect, in a possible implementation manner, the problems in the historical session content include artificial intelligence acceleration device installation problems and / or software deployment problems related to the artificial intelligence acceleration device.
[0021] In the above embodiments, the conversation system is used to automatically solve common problems encountered in the installation of artificial intelligence acceleration devices and / or the deployment of software related to artificial intelligence acceleration devices, which can improve efficiency and reduce manual workload.
[0022] In a second aspect, an embodiment of the present application provides a conversation method, which is applied to a conversation management module, and the method includes: obtaining a first query request sent by an agent based on the current conversation content of a customer; querying a database for target historical conversation content related to the first query request to obtain a first query result, wherein the database stores historical conversation content between different agents and customers, and the historical conversation content includes questions raised by the customer and solutions given by the agent to the questions; sending the first query result to the agent, wherein when the first query result represents a query to the target historical conversation content, the agent obtains a first solution based on the target historical conversation content and the current conversation content.
[0023] In a third aspect, an embodiment of the present application provides a conversation method, applied to an intelligent agent, the method comprising: sending a first query request to a conversation management module based on a current conversation content of a customer, wherein the first query request is used to query a database for target historical conversation content related to the first query request; the database stores historical conversation content between different intelligent agents and customers, the historical conversation content including questions raised by the customer and solutions given by the intelligent agent to the questions; receiving a first query result returned by the conversation management module in response to the first query request; and when the first query result represents a query to the target historical conversation content, obtaining a first solution based on the target historical conversation content and the current conversation content.
[0024] In a fourth aspect, an embodiment of the present application provides an electronic device, comprising: a memory and a processor, the processor being connected to the memory; the memory being used to store programs; and the processor being used to call the programs stored in the memory to execute the method provided in the above-mentioned second aspect embodiment or the third aspect embodiment.
[0025] In a fifth aspect, an embodiment of the present application provides a computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the method provided by the above-mentioned second aspect embodiment or third aspect embodiment is executed.
[0026] Other features and advantages of the present application will be described in the following description. The purpose and other advantages of the present application can be realized and obtained through the structures specifically pointed out in the written description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0027] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings. The above and other purposes, features and advantages of the present application will be more clearly shown in the drawings.
[0028] Figure 1 A structural diagram of a conversation system provided in an embodiment of the present application is shown.
[0029] Figure 2 A flow chart of a conversation method provided in an embodiment of the present application is shown.
[0030] Figure 3 A flow chart of another conversation method provided in an embodiment of the present application is shown.
[0031] Figure 4A schematic diagram showing the principles of a conversation method provided in an embodiment of the present application is shown.
[0032] Figure 5 A schematic structural diagram of an electronic device provided in an embodiment of the present application is shown. DETAILED DESCRIPTION
[0033] The technical solutions in the embodiments of the present application will be described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. The following embodiments can be used as examples to more clearly illustrate the technical solutions of the present application, and cannot be used to limit the scope of protection of the present application. It will be appreciated by those skilled in the art that, in the absence of conflict, the features in the following embodiments and embodiments can be combined with each other.
[0034] It should be noted that similar numbers and letters represent similar items in the following figures, so once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. At the same time, in the description of this application, relational terms such as "first", "second", etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "comprise", "include" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.
[0035] Furthermore, the term "and / or" in this application is merely a term used to describe the association relationship between associated objects, indicating that three relationships may exist. For example, A and / or B may represent three situations: A exists alone, A and B exist at the same time, and B exists alone.
[0036] In the description of the embodiments of the present application, unless otherwise clearly specified or limited, the technical term "connection" may refer to a direct connection or an indirect connection through an intermediate medium.
[0037] An embodiment of the present application provides a conversation system, which sets a conversation management module so that an intelligent agent can query historical conversation content related to the query request when answering customer questions, thereby enabling the intelligent agents to share conversation content, breaking down the barriers of "knowledge islands" between intelligent agents, and allowing different intelligent agents to give consistent solutions to similar problems, which not only improves the customer experience, but also improves the utilization rate of knowledge. In addition, by sharing conversation content, the intelligent agent can better understand the questions raised by the customer and give more comprehensive and accurate solutions, thereby providing more accurate decisions and improving the quality of decisions.
[0038] like Figure 1 As shown, Figure 1 A structural schematic diagram of a conversation system provided by an embodiment of the present application is shown. The conversation system may include a conversation management module and multiple agents, each of which is connected to the conversation management module. The conversation management module manages the historical conversation content between different agents, so that the conversation content is shared between different agents, breaking the barrier of "knowledge islands" and improving the utilization rate of knowledge.
[0039] The session management module is used to query the target historical session content related to the first query request from the database. After receiving the query request from the intelligent agent, the session management module will query the target historical session content related to the query request from the database.
[0040] In a possible implementation, the session management module may be another module independent of the agent. In a possible implementation, the session management module may be an agent.
[0041] The database stores historical conversation content between different intelligent agents and customers, including questions raised by customers and solutions given by the intelligent agents to the problems.
[0042] Among them, after receiving the historical conversation content sent by different intelligent agents, the conversation management module will save it in the database. When saving the historical conversation content, the conversation management module will first determine whether there is similar historical conversation content similar to the historical conversation content in the database. In one implementation, if there is similar historical content related to the historical conversation content in the database, the conversation management module will merge the historical conversation content with the similar historical conversation content and then save it in the database. In one implementation, if there is no similar historical content related to the historical conversation content in the database, the conversation management module directly saves the historical conversation content in the database.
[0043] In one possible implementation, the database is located in the session management module, or is independent of the session management module. In one possible implementation, the database may exist as an independent component of the session system. In some possible implementations, the database may not belong to the session system, and the session system may be connected to the database.
[0044] The problems in the above historical conversation content include, but are not limited to, customer service and support, online education and learning, retail and e-commerce, etc. For example, when the problem in the historical conversation content is a customer support problem, the customer support problem may include an AI acceleration device installation problem and / or an AI acceleration device-related software deployment problem. For example, system environment incompatibility, missing dependency libraries, environment variable configuration problems, etc.
[0045] Among them, AI acceleration devices refer to hardware devices specially designed to improve the performance of AI (Artificial Intelligence) applications. These devices accelerate the training and reasoning process of machine learning (ML) models by providing specialized computing power. For example, AI acceleration cards, ASICs (Application-Specific Integrated Circuits), GPUs (Graphics Processing Units), etc.
[0046] The above-mentioned software deployment issues related to artificial intelligence acceleration devices may include SDK (Software Development Kit) deployment issues. SDK deployment refers to the installation and configuration process of the software development kit in actual applications.
[0047] In a possible implementation, the session management module is further configured to, when receiving multiple similar second query requests in a target time period, fuse the query contents in the multiple second query requests to obtain a target query request; query the target historical session content related to the target query request from a database to obtain a second query result; and send the second query results to the agents that sent the second query requests. After receiving multiple similar second query requests from different agents in a target time period, the session management module fuses the query contents of the multiple similar second query requests, and then query the target historical session content related to the fused query request from a database based on the fused query request to obtain a second query result; and then send the second query results to the agents that sent the second query requests.
[0048] The session management module may determine, through the central processing unit, whether multiple requests received in the target time period are similar.
[0049] The above embodiment will be described below by using an example.
[0050] The query content of request 1 is: the power supply of device K of model Y is fine, but it cannot be started. The query content of request 2 is: the device K of model Y has been fully charged, but the screen of device K is not lit, and after trying to restart, the screen is still not lit. The session management module analyzes the content of request 1 and the content of request 2, and recognizes that request 1 is similar to request 2, which may point to a common fault of the same batch of devices. The session management module merges the content of request 1 and request 2 to obtain request 3: the power supply of device K of model Y is fine, but it cannot be started; it has been fully charged, but the screen of device K is not lit, and after trying to restart, the screen is still not lit. The session management module queries the target historical session content related to request 3 from the database according to request 3, and obtains the second query result, which may include that device K of model Y needs to be updated with firmware. The session management module sends the second query result to the intelligent agent that sends request 1 and the intelligent agent that sends request 2, respectively.
[0051] The intelligent agent is used to send a first query request to the session management module according to the customer's current session content, and receive a first query result returned by the session management module in response to the first query request. When the first query result represents the query to the target historical session content, a first solution is obtained based on the target historical session content and the current session content.
[0052] There are multiple agents in the conversation system, and multiple agents can independently conduct conversations with customers. Each agent can be regarded as a system with brain-like, perception and action capabilities. The brain part, as a controller, undertakes tasks such as memory, thinking, reasoning and decision-making; the perception part is responsible for perceiving and processing multimodal information from the external environment; the action part is responsible for using tools to perform tasks and affect the surrounding environment. The components of each agent can be divided into four components: planning, memory, action and tools according to their functions and roles. The planning component has functions such as goal setting, strategy generation, decision making, prediction and optimization. The memory component has functions such as information storage, knowledge accumulation, context understanding, and state tracking. The tool component provides the resources and tools required to realize the functions of the agent and assists the agent in making decisions. According to the instructions of the planning component, the action component mobilizes memory and tool resources to accurately and accurately perform each step of the operation.
[0053] In one possible implementation, the agent may include a large language model. The specific process of the agent obtaining the first solution based on the target historical conversation content and the current conversation content may be: inputting the target historical conversation content and the current conversation content into the large language model, and the large language model outputs the first solution.
[0054] The large language model in each agent can be a private domain large language model, that is, a model that adjusts the general model based on the conversation system using the enterprise's private knowledge base. The general large language model can be ChatGPT, LLaMA2, DALL-E, etc.
[0055] In the application process of the conversation system provided by the embodiment of the present application, the staff can work with the intelligent agent to solve the customer's problem. Through the intervention of the staff, not only can a more personalized and empathetic interaction be provided to increase customer satisfaction, but also customer feedback can be collected in a timely manner for service improvement.
[0056] The above embodiment is described below by an example. In the case where the first query result returned by the session management module in response to the first query request received by the agent represents the query of the target historical session content related to the first query request, the process of the agent and the staff collaborating to solve the customer problem can be shown in Table 1.
[0057]
[0058] In a possible implementation, the agent is also used to determine the type of question in the current conversation content; different thinking modes are selected according to the type of question, for example, if the type of question in the current conversation content is a complex question, the agent enters a slow thinking mode; if the type of question in the current conversation content is a simple question, the agent enters a fast thinking mode.
[0059] Among them, complex problems can be problems with complex external dependencies such as network fault diagnosis, hardware selection and compatibility, and software environment construction. Simple problems can be problems such as checking account balances and retrieving passwords. In one possible implementation, the agent can use an NLP (Natural Language Processing) classification algorithm to determine the type of question in the current conversation content. Among them, the NLP classification algorithm can assign text data to predefined categories or labels. The NLP classification algorithm includes but is not limited to support vector machines, random forests, convolutional neural networks, recurrent neural networks and other algorithms.
[0060] In the above embodiment, entering the slow thinking mode may include: sending a first query request to the session management module based on the customer's current session content, and receiving a first query result returned by the session management module in response to the first query request, and when the first query result represents that the target historical session content has been queried, obtaining a first solution based on the target historical session content and the current session content.
[0061] By adopting the slow thinking mode, the intelligent agent can refer to and integrate historical conversation content, thereby enabling the intelligent agent to more accurately grasp the customer's problems and make more comprehensive decisions, thereby improving the quality of decision-making and the customer's experience.
[0062] Among them, the fast thinking mode includes: getting the second solution directly based on the current conversation content. Since the fast thinking mode does not involve the query and fusion of shared conversation content between different agents, it can enable the agent to respond to customer questions quickly and improve work efficiency; at the same time, for simple problems, the fast thinking mode can reduce the consumption of computing resources and save resources.
[0063] In one possible implementation, the intelligent agent is also used to obtain the customer's first response based on the second solution; when the first response indicates that the problem in the current conversation content is not resolved, the slow thinking mode is entered.
[0064] When the fast thinking mode cannot solve the customer's problem, we will switch to the slow thinking mode in time, and provide more comprehensive and accurate solutions through in-depth analysis of the customer's problem and reference to relevant historical conversation content, thereby improving the customer experience and satisfaction.
[0065] The above embodiment will be described below by an example. When the type of question in the current conversation content is a simple question, after entering the fast thinking mode, the customer's question is not solved, and the slow thinking mode is switched, the process of the intelligent agent solving the customer's question can be shown in Table 2.
[0066]
[0067]
[0068] In one possible implementation, the intelligent agent may include a dynamic switching controller, a rapid response unit, and a deep analysis unit.
[0069] The dynamic switching controller is used to dynamically switch to the fast thinking mode or the slow thinking mode according to the type of question in the current conversation content.
[0070] The deep analysis unit is used to send a first query request to the session management module according to the current session content of the client, and receive a first query result returned by the session management module in response to the first query request, and when the first query result indicates that the target historical session content has been queried, a first solution is obtained based on the target historical session content and the current session content. That is, the deep analysis unit is used to execute the slow thinking mode.
[0071] The quick response unit is used to obtain a second solution directly based on the current conversation content. That is, the quick response unit is used to execute the quick thinking mode.
[0072] In one possible implementation, the intelligent agent is also used to desensitize the content of the current session after the current session ends, obtain the first historical session content including the problem and the solution given by the intelligent agent to the problem, and send the first historical session content to the session management module for storage.
[0073] Among them, desensitization is a security technology used to protect private data and sensitive information. It deforms sensitive data to achieve reliable protection of sensitive privacy data. Desensitization methods can include replacement, randomization, overwriting, reversibility, reversible restoration, truncation, interception, mapping, encryption, etc.
[0074] The above embodiment will be described below by an example. The content of this session is shown in Table 3.
[0075]
[0076] After desensitizing the content of this conversation, the first historical conversation content including the problem and the solution given by the intelligent agent to the problem can be shown in Table 4.
[0077]
[0078]
[0079] In this implementation, the session management module is further used to query the database for similar historical session content related to the first historical session content when receiving the first historical session content; when similar historical session content is found, merge the first historical session content and the similar historical session content to obtain merged historical session content; and save the merged historical session content to the database. In one implementation, when similar historical session content is not found, the first historical session content is directly saved to the database.
[0080] By updating the conversation content to the database in a timely manner, the intelligent agent can obtain the latest historical conversation content, which helps the intelligent agent provide more accurate and comprehensive solutions; and timely updating of the database can provide the intelligent agent with richer data resources, which helps improve its ability to handle complex problems; in addition, timely updating of the database can enable different intelligent agents to give consistent solutions when facing similar problems, thereby improving customer experience.
[0081] In a possible implementation, the conversation system provided in the embodiment of the present application may further include a knowledge base. In this implementation, the conversation management module is further used to update the knowledge base using the first historical conversation content or the merged historical conversation content.
[0082] The knowledge base is the core component of a conversational system for understanding and answering customer questions. It is a collection of data containing a large amount of information that is organized and stored so that the conversational system can quickly retrieve and use it.
[0083] like Figure 2 As shown, Figure 2 A flow chart of a conversation method provided in an embodiment of the present application is shown. The conversation method is applied to a conversation management module. The conversation method may include the following steps:
[0084] Step S100: Obtain the first query request sent by the agent according to the current session content of the client.
[0085] Step S200: querying the database for target historical conversation content related to the first query request to obtain a first query result.
[0086] The database stores historical conversation content between different intelligent agents and customers, including questions raised by customers and solutions given by the intelligent agents to the problems.
[0087] Step S300: Send the first query result to the agent.
[0088] When the first query result represents that the target historical conversation content is queried, the agent obtains a first solution based on the target historical conversation content and the current conversation content.
[0089] In one possible implementation, the conversation method may also include: when multiple similar second query requests are received in a target time period, fusing the query contents in the multiple second query requests to obtain a target query request; querying the target historical conversation content related to the target query request from a database to obtain a second query result; and sending the second query results to the intelligent agents that sent the second query requests respectively.
[0090] In one possible implementation, the conversation method may also include: receiving a first historical conversation content sent by an intelligent agent, and querying a database for similar historical conversation content related to the first historical conversation content; when similar historical conversation content is found, merging the first historical conversation content and the similar historical conversation content to obtain merged historical conversation content; and saving the merged historical conversation content to a database.
[0091] Among them, the first historical conversation content is obtained by the intelligent agent performing desensitization processing on the content of this conversation.
[0092] In a possible implementation manner, the conversation method may further include: when no similar historical conversation content is found, directly saving the first historical conversation content into a database.
[0093] The conversation method provided in the embodiment of the present application has the same implementation principle and technical effects as those of the aforementioned system embodiment. For the sake of brief description, for matters not mentioned in the method embodiment, reference may be made to the corresponding contents in the aforementioned system embodiment.
[0094] like Figure 3 As shown, Figure 3 A flow chart of another conversation method provided in an embodiment of the present application is shown. The conversation method is applied to an agent. The conversation method may include the following steps:
[0095] Step S10: Sending a first query request to the session management module according to the current session content of the client.
[0096] Among them, the first query request is used to query the target historical conversation content related to the first query request from the database; the database stores historical conversation content between different intelligent agents and customers, and the historical conversation content includes questions raised by customers and solutions given by the intelligent agents to the problems.
[0097] Step S20: Receive a first query result returned by the session management module in response to the first query request.
[0098] Step S30: when the first query result indicates that the target historical conversation content is found, a first solution is obtained based on the target historical conversation content and the current conversation content.
[0099] In a possible implementation manner, the conversation method may further include: determining the type of question in the current conversation content; and entering a slow thinking mode when the type of question in the current conversation content is a complex question.
[0100] Among them, entering the slow thinking mode includes: sending a first query request to the session management module according to the customer's current session content, and receiving a first query result returned by the session management module in response to the first query request, and when the first query result represents that the target historical session content is queried, a first solution is obtained based on the target historical session content and the current session content.
[0101] In a possible implementation, the conversation method may further include: when the type of question in the current conversation content is a simple question, entering a quick thinking mode.
[0102] Among them, the quick thinking mode includes: obtaining a second solution directly based on the current conversation content.
[0103] In a possible implementation manner, the conversation method may further include: obtaining a first response from the customer based on the second solution; and entering a slow thinking mode when the first response indicates that a problem in the current conversation content is not resolved.
[0104] In one possible implementation, the conversation method may also include: after the current conversation ends, desensitizing the content of the current conversation to obtain a first historical conversation content including the problem and the solution given by the intelligent agent to the problem, and sending the first historical conversation content to the conversation management module for storage.
[0105] The following will be combined Figure 4 The principle schematic diagram shown illustrates the principle of the conversation method applied to the agent through an example. Determine the type of problem in the current conversation content. When the problem is a complex problem, enter the slow thinking mode, that is, send a first query request to the conversation management module according to the customer's current conversation content, and receive the first query result returned by the conversation management module in response to the first query request; when the first query result represents the query of the target historical conversation content, obtain the first solution based on the target historical conversation content and the current conversation content. Obtain the second response of the customer based on the first solution, and determine whether the second response represents that the problem in the current conversation content has been solved. If the second response represents that the problem in the current conversation content has not been solved, apologize to the customer and record the problem in the current conversation content for subsequent problem analysis and system optimization. If the agent and the staff work together to solve the customer's problem, when the second response represents that the problem in the current conversation content has not been solved, the agent can automatically transfer to manual service, or provide the customer with the option of manual service. If the second response indicates that the problem in the current conversation content has been solved, the conversation is ended, and the content of this conversation is desensitized to obtain the first historical conversation content including the problem and the solution given by the intelligent agent to the problem, and the first historical conversation content is sent to the conversation management module for storage.
[0106] When the problem is a simple one, enter the fast thinking mode, that is, get the second solution directly based on the current session content. Get the customer's first response based on the second solution, and determine whether the first response indicates that the problem in the current session content has been solved. If the first response indicates that the problem in the current session content has not been solved, enter the slow thinking mode. If the first response indicates that the problem in the current session content has been solved, end the session, desensitize the content of this session, obtain the first historical session content containing the problem and the solution given by the intelligent agent to the problem, and send the first historical session content to the session management module for storage.
[0107] The conversation method provided in the embodiment of the present application has the same implementation principle and technical effects as those of the aforementioned system embodiment. For the sake of brief description, for matters not mentioned in the method embodiment, reference may be made to the corresponding contents in the aforementioned system embodiment.
[0108] like Figure 5 As shown, Figure 5The schematic diagram of the structure of an electronic device 20 provided in an embodiment of the present application is shown. The electronic device 20 includes: a processor 21 and a memory 22.
[0109] Figure 5 The components and structures of the electronic device 20 shown are merely exemplary and non-limiting. The electronic device 20 may also have other components and structures as required.
[0110] The processor 21, the memory 22, and other components that may appear in the electronic device 20 are directly or indirectly electrically connected to each other to achieve data transmission or interaction. For example, the processor 21, the memory 22, and other components that may appear may be electrically connected to each other via one or more communication buses or signal lines.
[0111] The memory 22 is used to store programs, for example, the program corresponding to the conversation method mentioned above is stored.
[0112] The processor 21 is used to call the program stored in the memory to execute the conversation method mentioned above.
[0113] Among them, the memory 22 can be, but is not limited to, random access memory (Random Access Memory, RAM), read only memory (Read Only Memory, ROM), programmable read-only memory (Programmable Read-Only Memory, PROM), erasable programmable read-only memory (Erasable Programmable Read-Only Memory, EPROM), electrically erasable read-only memory (Electric Erasable Programmable Read-Only Memory, EEPROM), etc.
[0114] The processor 21 may be an integrated circuit chip with signal processing capabilities. The above-mentioned processor 21 may be a general-purpose processor, including a central processing unit (CPU), a network processor (NP), a graphics processor (GPU), an accelerated processing unit (Accelerated Processing Unit), a multimedia application processor (MAP), a microprocessor, etc.; it may also be a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components. The disclosed methods, steps and logic block diagrams in the embodiments of the present application can be implemented or executed. Or the processor 21 may also be any conventional processor, etc.
[0115] The electronic device 20 mentioned above includes but is not limited to a computer, a mobile phone, an iPad, a server, a robot, etc. In some possible implementations, the electronic device 20 may be an electronic device including a GPU graphics card.
[0116] An embodiment of the present application also provides a non-volatile computer-readable storage medium (hereinafter referred to as storage medium), on which a computer program is stored. When the computer program is run by a computer such as the above-mentioned electronic device 20, the method disclosed in any of the above-mentioned embodiments is executed.
[0117] It should be noted that the various embodiments in this specification are described in a progressive manner, and each embodiment focuses on the differences from other embodiments. The same or similar parts between the various embodiments can be referenced to each other.
[0118] In several embodiments provided in the present application, it should be understood that the disclosed devices and methods can also be implemented in other ways. The device embodiments described above are merely schematic. For example, the flowcharts and block diagrams in the accompanying drawings show the possible architecture, functions and operations of the devices, methods and computer program products according to multiple embodiments of the present application. In this regard, each box in the flowchart or block diagram can represent a module, a program segment or a part of a code, and the module, a program segment or a part of a code contains one or more executable instructions for implementing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two consecutive boxes can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram and / or flowchart, and the combination of boxes in the block diagram and / or flowchart can be implemented with a dedicated hardware-based system that performs a specified function or action, or can be implemented with a combination of dedicated hardware and computer instructions.
[0119] In addition, the functional modules in the various embodiments of the present application may be integrated together to form an independent part, or each module may exist separately, or two or more modules may be integrated to form an independent part.
[0120] If the functions are implemented in the form of software function modules and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application can be essentially or partly embodied in the form of a software product that contributes to the prior art. The computer software product is stored in a computer-readable storage medium, including several instructions for enabling a computer device (which can be a personal computer, a laptop, a server, or an electronic device, etc.) to perform all or part of the steps of the methods described in each embodiment of the present application. The aforementioned computer-readable storage medium includes: various media that can store program codes, such as a USB flash drive, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk.
[0121] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.
Claims
1. A conversation system, characterized in that: include: A conversation management module, configured to query a database for target historical conversation content related to the first query request, wherein the database stores historical conversation content between different agents and customers, and the historical conversation content includes questions raised by the customers and solutions given by the agents to the questions; The intelligent agent is used to send the first query request to the session management module according to the customer's current session content, and receive the first query result returned by the session management module in response to the first query request, and when the first query result represents the query to the target historical session content, obtain the first solution based on the target historical session content and the current session content.
2. The conversation system according to claim 1, characterized in that: The intelligent agent is further used to determine the type of question in the current conversation content; When the type of question in the current conversation content is a complex question, entering a slow thinking mode; Among them, entering the slow thinking mode includes: sending the first query request to the session management module according to the customer's current session content, and receiving the first query result returned by the session management module in response to the first query request, when the first query result represents the query to the target historical session content, obtaining a first solution based on the target historical session content and the current session content.
3. The conversation system according to claim 2, characterized in that: The intelligent agent is also used to enter a quick thinking mode when the type of question in the current conversation content is a simple question; wherein the quick thinking mode includes: obtaining a second solution directly based on the current conversation content.
4. The conversation system according to claim 3, characterized in that: The intelligent agent is further used to obtain a first response from the customer based on the second solution; When the first response indicates that the problem in the current conversation content is not resolved, the slow thinking mode is entered.
5. The conversation system according to claim 1, characterized in that: The intelligent agent is also used to desensitize the content of the current session after the current session ends, obtain the first historical session content including the problem and the solution given by the intelligent agent to the problem, and send the first historical session content to the session management module for storage.
6. The conversation system according to claim 5, characterized in that: The session management module is further configured to query the database for similar historical session content related to the first historical session content upon receiving the first historical session content; When the similar historical conversation content is found, merging the first historical conversation content with the similar historical conversation content to obtain merged historical conversation content; The merged historical conversation content is saved in the database.
7. The conversation system according to claim 1, characterized in that: The session management module is further configured to, when receiving a plurality of similar second query requests in a target time period, merge the query contents in the plurality of second query requests to obtain a target query request; Querying the database for target historical conversation content related to the target query request to obtain a second query result; The second query results are respectively sent to the agent that sent the second query request.
8. The conversation system according to claim 1, characterized in that: The issues in the historical session content include AI acceleration device installation issues and / or AI acceleration device-related software deployment issues.
9. A conversation method, characterized in that: Applied to a session management module, the method comprises: Obtaining a first query request sent by the agent according to the current session content of the client; Querying a database for target historical conversation content related to the first query request to obtain a first query result, wherein the database stores historical conversation content between different agents and customers, and the historical conversation content includes questions raised by the customers and solutions given by the agents to the questions; The first query result is sent to the agent, wherein when the first query result represents that the target historical conversation content is queried, the agent obtains a first solution based on the target historical conversation content and the current conversation content.
10. A conversation method, characterized in that: Applied to an intelligent agent, the method comprises: Sending a first query request to the session management module according to the current session content of the customer, wherein the first query request is used to query the target historical session content related to the first query request from a database; the database stores historical session content between different intelligent agents and customers, and the historical session content includes questions raised by the customer and solutions given by the intelligent agent to the questions; Receiving a first query result returned by the session management module in response to the first query request; When the first query result indicates that the target historical conversation content is queried, a first solution is obtained based on the target historical conversation content and the current conversation content.
11. An electronic device, characterized in that: include: A memory and a processor, wherein the processor is connected to the memory; The memory is used to store programs; The processor is used to call the program stored in the memory to execute the method according to claim 9, or to execute the method according to claim 10.
12. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the computer program is executed by a processor, the method according to claim 9 is executed, or the method according to claim 10 is executed.