User intention conversion processing method and apparatus, computer device, and readable storage medium

Abstract

The application provides a user intention conversion processing method and device, computer equipment and a readable storage medium, and relates to the field of artificial intelligence. The method comprises the following steps: receiving a first session request of a user, wherein the first session request at least carries a call identifier and historical dialogue content; generating a structured session context based on the historical dialogue content, and generating a robot main reply result based on the structured session context; starting an asynchronous intention recognition process relative to the generated robot main reply result to perform intention recognition on several rounds of dialogue of the user, and obtaining an intention recognition result; if the intention recognition result is that the user has a conversion intention, writing the intention recognition result into a state cache, and associating the intention recognition result with the call identifier in the state cache; receiving a second session request of the user, wherein the second session request at least carries the call identifier; and when it is determined based on the call identifier that the intention recognition result is that the user has a conversion intention, performing a user conversion action.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence, and in particular to a user intention conversion processing method, apparatus, computer device, and readable storage medium. Background Technology

[0002] With the development of artificial intelligence, automatic speech recognition, and large-scale dialogue technology, outbound call robots have been widely used in industries such as insurance, education, healthcare, finance, and customer service. In these scenarios, the call robot not only needs to engage in natural conversation with the user but also needs to identify the user's intent. Most existing call robot systems currently integrate the generation of the main dialogue and the determination of user intent into the same processing chain. While generating the response, the robot also needs to simultaneously determine whether the user intends to continue communication.

[0003] The existing technologies mentioned above have the following main problems: (1) The main dialogue link and the user intent judgment link are heavily coupled. When the robot is responsible for generating natural dialogue and completing the conversion judgment in the same processing link, it is easy to cause the interface response to slow down and affect the real-time interactive experience during the outbound call process. (2) Due to the impact on real-time interaction, user replies in the telephone conversation often have short pauses, interruptions, repeated recognition and recognition frame supplementation. If the user intent judgment must be completed synchronously, it is easy to cause misjudgment, missed judgment or block the main reply. Summary of the Invention

[0004] In view of this, this application provides a user intention conversion processing method, apparatus, computer device, and readable storage medium, which solves the above-mentioned technical problems.

[0005] In a first aspect, embodiments of this application provide a user intent conversion processing method, including: Receive the user's first session request, which carries at least a call identifier and historical conversation content; A structured conversation context is generated based on the historical dialogue content, and the main response result of the robot is generated based on the structured conversation context; An asynchronous intention recognition process relative to the generation of the robot's main response result is initiated to perform intention recognition on several rounds of dialogue with the user and obtain intention recognition results; If the intention recognition result indicates that the user has the intention to convert, then the intention recognition result is written into the state cache, and the intention recognition result is associated with the call identifier in the state cache; Receive the user's second session request, wherein the second session request carries at least the call identifier; When the intention recognition result is determined based on the call identifier to indicate that the user has the intention to convert, the user conversion action is executed.

[0006] Secondly, embodiments of this application provide a user intent conversion processing apparatus, including: The session access module is used to receive the user's first session request, which carries at least a call identifier and historical dialogue content. The main response generation module is used to generate a structured conversation context based on the historical dialogue content, and to generate the robot's main response result based on the structured conversation context; An asynchronous intention recognition module is used to initiate an asynchronous intention recognition process relative to the generation of the robot's main response result, so as to perform intention recognition on several rounds of dialogue of the user and obtain intention recognition results; The state caching module is used to write the intention recognition result into the state cache if the intention recognition result indicates that the user has the intention to convert, and to associate the intention recognition result with the call identifier in the state cache; The session access module is further configured to receive a second session request from the user, wherein the second session request carries at least the call identifier; The action output module is used to execute a user conversion action when the intention recognition result is determined based on the call identifier to indicate that the user has the intention to convert.

[0007] Thirdly, embodiments of this application provide a computer device including a processor and a memory, the memory storing programs or instructions that can run on the processor, the programs or instructions implementing the steps of the method as described in the first aspect when executed by the processor.

[0008] Fourthly, embodiments of this application provide a readable storage medium on which a program or instructions are stored, which, when executed by a processor, implement the steps of the method as described in the first aspect.

[0009] The user intent conversion processing method, apparatus, computer device, and readable storage medium of this application, in a telephone outbound call scenario, asynchronously identify user conversion intent and perform user conversion actions while ensuring real-time response of the main dialogue quasi-link. In this application, the main response link and the conversion judgment link are asynchronously decoupled, improving the real-time response capability of the telephone outbound robot. Furthermore, since the judgment of user conversion intent does not affect the main call interaction, there will be no pauses, interruptions, repeated recognition, or recognition frame supplementation, avoiding problems such as misjudgment, missed judgment, or blocking of the main response, thus improving the execution stability and business adaptability of the telephone robot's conversion actions.

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

[0011] The accompanying drawings, which are included to provide a further understanding of this application and form part of this application, illustrate exemplary embodiments and are used to explain this application, but do not constitute an undue limitation of this application. In the drawings: Figure 1 One of the flowcharts illustrating a user intent conversion processing method according to an embodiment of this application is shown; Figure 2 This is a second schematic flowchart illustrating the user intent conversion processing method according to an embodiment of this application; Figure 3 A structural block diagram of a user intent conversion processing apparatus according to an embodiment of this application is shown; Figure 4 A system architecture diagram of the user intent conversion processing apparatus according to an embodiment of this application is shown; Figure 5 A structural block diagram of a computer according to an embodiment of this application is shown. Detailed Implementation

[0012] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0013] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such use of data can be interchanged where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0014] The user intent conversion processing method, apparatus, computer equipment, and readable storage medium provided in this application will be described in detail below with reference to the accompanying drawings and through specific embodiments and application scenarios. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0015] This application provides a user intent conversion processing method, applied to a telephone outbound call robot. For example... Figure 1 and Figure 2 As shown, the method includes: Step 101: Receive the user's first session request, which includes at least a call identifier and historical conversation content.

[0016] In this step, the user's first session request is received. The call identifier and historical dialogue content are extracted from the first session request. User attribute information can also be extracted from the first session request. Historical dialogue content refers to all dialogue content completed between the outbound call robot and the user from the start of the current call until the initiation of this session request. This content belongs only to historical interactions under the same call identifier and does not cross calls or sessions.

[0017] The system cleans and structures historical dialogue content, and combines this with the speech recognition status to determine whether the current dialogue turn is new. If the speech recognition status indicates completion, it is considered a new dialogue turn; otherwise, it is considered a continuation of the same turn, not a new turn. This distinguishes between intermediate recognition results and complete user statements, avoiding repeated understanding and generation of responses for the same user input, ensuring correct dialogue sequence, clear context, and logical turn order, thus guaranteeing smooth and stable dialogue.

[0018] Step 102: Generate a structured conversation context based on the historical dialogue content, and generate the robot's main response result based on the structured conversation context.

[0019] In this step, a structured conversation context is generated based on historical dialogue content, call identifiers, and user attributes. Using this structured conversation context as input, the current user's semantics are parsed and combined with multi-turn historical dialogue context understanding. A rule-based dialogue engine or large-scale model is used to match the corresponding scenario's dialogue and response logic, quickly generating the robot's main response content adapted to the current conversation scenario. This ensures real-time response in the main dialogue chain, and the main response result is returned before the conversion action judgment result.

[0020] Step 103: Initiate an asynchronous intent recognition process relative to the generated robot main response result to perform intent recognition on several rounds of user dialogue and obtain intent recognition results.

[0021] In this step, while the robot prioritizes synchronously generating the main response based on the structured conversation context to ensure real-time dialogue response, an asynchronous intent recognition process is initiated separately. This process does not block the generation and return of the main response, but independently retrieves the user's multi-turn historical dialogue content for semantic analysis and intent judgment, and outputs the user intent recognition result after asynchronous processing, realizing parallel processing of real-time dialogue response and deep intent parsing.

[0022] In one embodiment of this application, an asynchronous intent recognition process is initiated relative to the generation of the robot's main response result to identify intents in several rounds of user dialogue and obtain intent recognition results, including: Determine whether the current round of dialogue meets the asynchronous intention recognition triggering conditions; If the conditions are met, an asynchronous intent recognition process relative to the generated robot main response result is initiated to perform intent recognition on several rounds of user dialogue and obtain intent recognition results.

[0023] The asynchronous intent recognition trigger conditions include: the existence of new dialogue content from the user, the fact that intent recognition has not been performed in the current round of dialogue, and the fact that user conversion actions have not been performed in the current round of dialogue.

[0024] In this embodiment, it is determined whether the current dialogue round meets the preset asynchronous intention recognition triggering conditions. Only when there is the user's latest dialogue content in the current round, the current round has not yet carried out intention recognition processing, and the current round has not performed any user conversion actions, will the asynchronous intention recognition process, which is independent of the main reply generation process, be started, so as to perform intention recognition on several rounds of the user's dialogue.

[0025] This asynchronous intent recognition process does not occupy the response time of the main reply, nor does it block the main session reply link. In the background, it performs semantic parsing and intent determination on the user's recent dialogue content in parallel, and finally outputs the corresponding intent recognition results. This ensures real-time response in human-computer dialogue and completes accurate analysis of user intent in the background.

[0026] Step 104: If the intent recognition result indicates that the user has the intention to convert, then write the intent recognition result into the state cache and associate the intent recognition result with the call identifier in the state cache.

[0027] In this step, if the intention identification determines that the user has the intention to convert, the intention identification result is written into the status cache. At the same time, the intention identification result is bound and associated with the current unique call identifier for storage, so that the intention data of different calls are isolated from each other and do not interfere with each other, making it easy for subsequent session requests to quickly read the corresponding user conversion intention information based on the call identifier.

[0028] In one embodiment of this application, the state cache also stores data on whether intention recognition has not been performed in the current round of dialogue, data on whether user conversion actions have been performed in the current round of dialogue, and the cumulative number of user conversion actions performed by the user.

[0029] In this embodiment, the asynchronous identification results, along with the corresponding round status, action status, and count information, are written into a state cache. The state cache records at least the following information: whether the user has been marked as a valid intent user, whether intent identification has been performed in the current round of dialogue, whether a conversion action has been returned in the current round of dialogue, and the cumulative number of user conversion actions. The cached state is associated with the call identifier and round information to accurately identify the current round processing status in subsequent requests.

[0030] Step 105: Receive the user's second session request, which contains at least a call identifier.

[0031] In this step, a second session request carrying a call identifier is received from the user, which means that the user has started a new round of conversation.

[0032] Step 106: When the intention recognition result based on the call identifier indicates that the user has the intention to convert, execute the user conversion action.

[0033] In this step, based on a new round of dialogue and once it is determined that the user has the intention to convert, the user conversion action is executed.

[0034] In related technologies, there is a lack of cross-round state memory capabilities. Many solutions only make a one-time judgment based on the current round of text, which cannot reliably record whether a user has been identified as a high-intent user, whether an intent detection has been performed in a certain round, and whether an action result has been returned in a certain round, thus leading to the problem of repeatedly executing conversion actions.

[0035] In this embodiment of the application, in a telephone outbound call scenario, while ensuring real-time response of the main dialogue link, the user's conversion intention is asynchronously identified, and a user conversion action is performed. In this application, the main response link and the conversion judgment link are asynchronously decoupled, improving the real-time response capability of the telephone outbound robot. Furthermore, since the judgment of user conversion intention does not affect the main call interaction, there will be no pauses, interruptions, repeated recognition, or recognition frame supplementation, avoiding problems such as misjudgment, missed judgment, or blocking of the main response, thus improving the execution stability and business adaptability of the telephone robot's conversion actions.

[0036] In related technologies, frequency control and deduplication control for business actions are lacking. When a CTI (Computer Telephony Integration) platform experiences duplicate callbacks, replays of ASR (Automatic Speech-to-Text) results, or multiple requests in the same round, the lack of round-level deduplication capabilities can easily lead to the same user being converted multiple times in a short period of time. In one embodiment of this application, executing a user conversion action includes: performing frequency control deduplication verification; if the verification passes, then executing the user conversion action.

[0037] In one embodiment of this application, frequency control deduplication verification is performed, including: Determine whether the user conversion action is repeated in the current round of dialogue, or whether the number of user conversion actions performed on the user exceeds a preset threshold; If the user conversion action is not repeated in the current round of dialogue, or the number of user conversion actions performed on the user does not exceed the preset threshold, then the verification is considered successful.

[0038] Frequency control refers to a control strategy that limits the number of times an action is triggered and the number of trigger rounds. Round deduplication refers to marking the status of recognition or actions that have already been performed within the same round of dialogue to prevent repeated execution.

[0039] In this embodiment, frequency control and deduplication checks are performed before the action is output to check whether the current round of dialogue repeats the conversion action and whether the number of conversion actions performed on the current user has reached a preset threshold. If the current round of dialogue does not repeat the user conversion action, or the number of user conversion actions performed on the user does not exceed the preset threshold, the check passes and the user conversion action is executed; otherwise, the repeated action is suppressed.

[0040] After the action is converted and output, the result of the action in the current round is written back to the state cache to prevent repeated triggering in subsequent rounds.

[0041] In related technologies, there is a lack of dynamic policy switching capabilities for different user attributes. The solutions in these technologies do not adequately utilize attributes such as whether adding an instant messaging account is allowed, whether the account has been successfully added, and whether an SMS message has been received, leading to inaccurate conversion actions. In one embodiment of this application, the first session request also carries user attribute information, which is used to characterize the policy type for executing the user conversion action; executing the user conversion action includes: If the strategy type is determined to be an instant messaging addition strategy based on user attribute information, then a user conversion action of adding an instant messaging account to the user is triggered. If the strategy type is determined to be SMS sending strategy based on user attribute information, then a user conversion action of sending SMS to the user will be triggered.

[0042] In this embodiment, an instant messaging addition strategy or an SMS sending strategy is selected by combining cache status and user attribute information. Specifically, the instant messaging addition strategy is output when the user meets the instant messaging addition conditions, and the SMS sending strategy is output when the user does not meet the instant messaging addition conditions but meets the SMS sending conditions.

[0043] It should be noted that, in addition to instant messaging and SMS, it can also be extended to other conversion actions such as email sending, app push notifications, manual transfer, and task creation.

[0044] In this embodiment, when the user's conversion intention is detected in a subsequent request (second session request), the system combines user attribute information, round status information, and action frequency information to explicitly select either an instant messaging addition strategy or an SMS sending strategy: if the user meets the instant messaging addition conditions, the instant messaging addition strategy is triggered; if the user does not meet the instant messaging addition conditions but meets the SMS sending conditions, the SMS sending strategy is triggered. If the user has already completed the corresponding action or the related action has reached the preset number of times limit, the system will not trigger the strategy again.

[0045] The embodiments of this application have the following technical effects: (1) The main response link and the conversion judgment link are asynchronously decoupled, which improves the real-time response capability of the telephone outbound robot.

[0046] (2) By using asynchronous intention recognition, the user's conversion intention can be continuously judged without affecting the main call interaction, thereby improving the system's throughput.

[0047] (3) By using the state caching mechanism, the user intention state, round recognition state and action return state can be shared across requests, thereby improving system stability.

[0048] (4) By dynamically switching between instant messaging and SMS sending strategies, appropriate conversion actions can be automatically matched according to different user attributes, thereby improving business adaptability.

[0049] (5) Through frequency control and deduplication mechanism, the problem of repeated conversion caused by repeated callbacks and multiple triggers in the same round can be effectively avoided, thus improving customer experience.

[0050] (6) This technical solution can be extended to multiple proactive outreach scenarios such as education, healthcare, finance, customer service follow-up, and course sales, and has good versatility and scalability.

[0051] In some embodiments, the solution of this application can not only be applied to user conversion in telephone marketing scenarios, but can also be extended to multiple proactive outreach scenarios such as education, healthcare, finance, customer service follow-up, and course sales, and has good versatility and scalability.

[0052] As a specific implementation of the above-mentioned user intention conversion processing method, this application provides a user intention conversion processing apparatus. For example... Figure 3 As shown, the user intent conversion processing device 300 includes: a session access module 301, a main response generation module 302, an asynchronous intent recognition module 303, a state cache module 304, and an action output module 305.

[0053] The session access module 301 is used to receive a user's first session request, which carries at least a call identifier and historical dialogue content. The main response generation module 302 is used to generate a structured conversation context based on the historical dialogue content, and to generate the robot's main response result based on the structured conversation context; The asynchronous intention recognition module 303 is used to initiate an asynchronous intention recognition process relative to the generation of the robot's main response result, so as to perform intention recognition on several rounds of dialogue of the user and obtain intention recognition results. The state caching module 304 is used to write the intention recognition result into the state cache if the intention recognition result indicates that the user has the intention to convert, and to associate the intention recognition result with the call identifier in the state cache; The session access module 301 is further configured to receive a second session request from the user, wherein the second session request carries at least the call identifier; The action output module 305 is used to execute a user conversion action when the intention recognition result is determined based on the call identifier to indicate that the user has the intention to convert.

[0054] Figure 4 A system architecture diagram of a user intention conversion processing device provided in an embodiment of this application is shown.

[0055] Furthermore, the asynchronous intention recognition module 303 is specifically used for: Determine whether the current round of dialogue meets the asynchronous intention recognition triggering conditions; If the conditions are met, an asynchronous intention recognition process relative to the generation of the robot's main response result is initiated to perform intention recognition on several rounds of dialogue with the user and obtain the intention recognition result.

[0056] Furthermore, the asynchronous intention recognition triggering conditions include: the existence of new dialogue content from the user, the absence of intention recognition in the current round of dialogue, and the absence of the user conversion action in the current round of dialogue.

[0057] Furthermore, the state cache also stores data on whether intention recognition has not been performed in the current round of dialogue, data on whether the user conversion action has been performed in the current round of dialogue, and the cumulative number of the user conversion actions performed by the user.

[0058] Furthermore, the motion output module 305 is specifically used for: Perform frequency control deduplication verification; If the verification passes, the user conversion action will be executed.

[0059] Furthermore, the motion output module 305 is specifically used for: Determine whether the user conversion action is repeated in the current round of dialogue, or whether the number of times the user conversion action is performed on the user exceeds a preset threshold; If the user conversion action is not repeated in the current round of dialogue, or if the number of times the user conversion action is performed on the user does not exceed a preset threshold, then the verification is deemed successful.

[0060] Furthermore, the first session request also carries user attribute information, which is used to characterize the strategy type for executing the user conversion action; the action output module 305 is specifically used for: If the strategy type is determined to be an instant messaging addition strategy based on the user attribute information, then a user conversion action of adding an instant messaging account to the user is triggered. If the strategy type is determined to be an SMS sending strategy based on the user attribute information, then a user conversion action of sending an SMS to the user is triggered.

[0061] The user intent conversion processing device 300 in this application embodiment can be a computer device or a component within a computer device, such as an integrated circuit or a chip. The user intent conversion processing device provided in this application embodiment can achieve... Figure 1 The various processes implemented in the user intent conversion processing method embodiment will not be described again here to avoid repetition.

[0062] This application embodiment can also be implemented as a system, which includes a telephone outbound call access system, a dialogue processing server, a caching service, a WeChat / Enterprise action execution service, an SMS action execution service, and a business management backend. The dialogue processing server integrates asynchronous intent recognition, state caching, strategy switching, and action frequency control logic, and outputs stable WeChat / SMS conversion results, equivalent to the user intent conversion processing device 300.

[0063] This application also provides a computer device, such as... Figure 5As shown, the computer device 500 includes a processor 501 and a memory 502. The memory 502 stores a program or instruction that can run on the processor 501. When the program or instruction is executed by the processor 501, it implements the various steps of the above-described user intention conversion processing method embodiment and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0064] The memory 502 can be used to store software programs and various data. The memory 502 may primarily include a first storage area for storing programs or instructions and a second storage area for storing data. The first storage area may store the operating system, application programs or instructions required for at least one function (such as sound playback, image playback, etc.). Furthermore, the memory 502 may include volatile memory or non-volatile memory, or both. The non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory can be random access memory (RAM), static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDRSDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchronous link dynamic random access memory (SLDRAM), and direct memory bus RAM (DRRAM). The memory 502 in this embodiment includes, but is not limited to, these and any other suitable types of memory.

[0065] Processor 501 may include one or more processing units; optionally, processor 501 integrates an application processor and a modem processor, wherein the application processor mainly handles operations involving the operating system, user interface, and applications, and the modem processor mainly handles wireless communication signals, such as a baseband processor. It is understood that the aforementioned modem processor may also not be integrated into processor 501.

[0066] This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described user intention conversion processing method embodiments and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0067] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes that element. Furthermore, it should be noted that the scope of the methods and apparatuses in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in the reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

[0068] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of this application without departing from the spirit and scope of the claims, and all of these forms are within the protection scope of this application.

Claims

1. A method for processing user intent conversion, characterized in that, include: Receive the user's first session request, which carries at least a call identifier and historical conversation content; A structured conversation context is generated based on the historical dialogue content, and the main response result of the robot is generated based on the structured conversation context; An asynchronous intention recognition process relative to the generation of the robot's main response result is initiated to perform intention recognition on several rounds of dialogue with the user and obtain intention recognition results; If the intention recognition result indicates that the user has the intention to convert, then the intention recognition result is written into the state cache, and the intention recognition result is associated with the call identifier in the state cache; Receive the user's second session request, wherein the second session request carries at least the call identifier; When the intention recognition result is determined based on the call identifier to indicate that the user has the intention to convert, the user conversion action is executed.

2. The method according to claim 1, characterized in that, The initiation of the asynchronous intent recognition process relative to the generation of the robot's main response result, to perform intent recognition on several rounds of dialogue with the user and obtain intent recognition results, includes: Determine whether the current round of dialogue meets the asynchronous intention recognition triggering conditions; If the conditions are met, an asynchronous intention recognition process relative to the generation of the robot's main response result is initiated to perform intention recognition on several rounds of dialogue with the user and obtain the intention recognition result.

3. The method according to claim 2, characterized in that, The asynchronous intention recognition triggering conditions include: the existence of new dialogue content from the user, the absence of intention recognition in the current round of dialogue, and the absence of the user conversion action in the current round of dialogue.

4. The method according to claim 1, characterized in that, The state cache also stores data on whether intention recognition has not been performed in the current round of dialogue, data on whether the user conversion action has been performed in the current round of dialogue, and the cumulative number of user conversion actions performed by the user.

5. The method according to claim 1, characterized in that, The execution of user conversion actions includes: Perform frequency control deduplication verification; If the verification passes, the user conversion action will be executed.

6. The method according to claim 5, characterized in that, The frequency control deduplication verification includes: Determine whether the user conversion action is repeated in the current round of dialogue, or whether the number of times the user conversion action is performed on the user exceeds a preset threshold; If the user conversion action is not repeated in the current round of dialogue, or if the number of times the user conversion action is performed on the user does not exceed a preset threshold, then the verification is deemed successful.

7. The method according to claim 1, characterized in that, The first session request also carries user attribute information, which is used to characterize the strategy type for executing the user conversion action; the execution of the user conversion action includes: If the strategy type is determined to be an instant messaging addition strategy based on the user attribute information, then a user conversion action of adding an instant messaging account to the user is triggered. If the strategy type is determined to be an SMS sending strategy based on the user attribute information, then a user conversion action of sending an SMS to the user is triggered.

8. A user intention conversion processing device, characterized in that, include: The session access module is used to receive the user's first session request, which carries at least a call identifier and historical dialogue content. The main response generation module is used to generate a structured conversation context based on the historical dialogue content, and to generate the robot's main response result based on the structured conversation context; An asynchronous intention recognition module is used to initiate an asynchronous intention recognition process relative to the generation of the robot's main response result, so as to perform intention recognition on several rounds of dialogue of the user and obtain intention recognition results; The state caching module is used to write the intention recognition result into the state cache if the intention recognition result indicates that the user has the intention to convert, and to associate the intention recognition result with the call identifier in the state cache; The session access module is further configured to receive a second session request from the user, wherein the second session request carries at least the call identifier; The action output module is used to execute a user conversion action when the intention recognition result is determined based on the call identifier to indicate that the user has the intention to convert.

9. A computer device, characterized in that, It includes a processor and a memory, the memory storing a program or instructions that run on the processor, the program or instructions being executed by the processor to implement the steps of the user intention conversion processing method as described in any one of claims 1 to 7.

10. A readable storage medium having a program or instructions stored thereon, characterized in that, When the program or instructions are executed by the processor, they implement the steps of the user intention conversion processing method as described in any one of claims 1 to 7.

Images