Ride-hailing driver and passenger negotiation assistance method and apparatus, electronic device, and storage medium

Abstract

The application discloses a network car-hailing driver and passenger negotiation assisting method and device, electronic equipment and a storage medium, wherein the method comprises the following steps: collecting basic information of a current order when a current user enters an IM dialogue interface, and the basic information comprises order identification information of the current order; when it is determined that there is historical dialogue information in the instant messaging dialogue interface according to the order identification information, obtaining a target text message, a sending client, a receiving client and a role type in the historical dialogue information, and the target text message is the last text message in the historical dialogue information; generating a rewritten text message based on the role type of the sending client, and sending the rewritten text message to the sending client, and / or generating a feedback text message of the target text message based on the role type of the receiving client and recommending the feedback text message to the receiving client. The application can improve the communication efficiency and user experience of the driver and the passenger, and improve the order success rate of the platform.

Description

Technical Field

[0001] This application belongs to the field of artificial intelligence, and in particular relates to a method, device, electronic device and storage medium for assisting in negotiation between ride-hailing drivers and passengers. Background Technology

[0002] With the rapid development of the ride-sharing industry, the real-time communication and negotiation between drivers (vehicle owners) and passengers on ride-hailing platforms such as carpooling has become a key factor affecting order completion rates. Current statistics show that among orders cancelled after acceptance, "failed negotiation" accounts for as much as 25%, making it one of the top three reasons for order cancellation. Among these, fee negotiations (such as highway toll negotiations and occupancy fee negotiations, accounting for 78%~85%) and travel time negotiations (accounting for 36%) are the core conflict scenarios. When negotiating key matters such as fee sharing and travel time adjustments, drivers and passengers often break down due to inappropriate wording and poor communication methods (e.g., from the passenger's perspective, "poor communication attitude" of drivers accounts for 18%), seriously affecting platform operational efficiency and user experience. Summary of the Invention

[0003] This application aims to address at least one of the technical problems existing in the related technologies. To this end, this application proposes a method, device, electronic device, and storage medium to assist in ride-hailing driver-passenger negotiation, thereby improving communication efficiency and user experience for both parties and increasing the platform's order success rate.

[0004] Firstly, this application provides a method for assisting in negotiation between ride-hailing drivers and passengers, applied to a server, the method comprising: Collect basic information about the current order when the current user enters the instant messaging chat interface. The basic information includes at least the order identifier information of the current order. When it is determined that there is historical dialogue information in the instant messaging interface based on the order identification information, the target text message, the sending client, the receiving client and the role type of the target text message in the historical dialogue information are obtained, and the target text message is the last text message in the historical dialogue information. A rewritten text message is generated based on the role type of the sending client, and the rewritten text message is sent to the sending client so that the sending client replaces the target text message with the rewritten text message, and / or a feedback text message of the target text message is generated based on the role type of the receiving client and recommended to the receiving client.

[0005] Secondly, this application provides a method for assisting in negotiation between ride-hailing drivers and passengers, applied to a client-side application. This method includes: When historical conversation information exists in the instant messaging interface of the client, the server receives and displays candidate rewritten text messages or candidate feedback text messages, wherein the candidate rewritten text messages and the candidate feedback text messages are generated by the server based on the role type of the client. In response to the current user's selection of a rewritten text message from the candidate rewritten text messages, the target text message is replaced with the rewritten text message, wherein the target text message is the last text message in the historical dialogue information; or In response to the current user's reply operation based on the candidate feedback text message, and based on the reply operation, perform the corresponding operation; When historical chat information exists in the instant messaging interface of the client, a default recommended text message sent by the server is received and displayed, wherein the default recommended text message is generated by the server based on the client's role type and order status information. In response to the target recommended text message selected by the current user in the default recommended text message, the target recommended text message is sent to the other party's client.

[0006] Thirdly, this application provides a ride-hailing driver-passenger negotiation assistance device, which includes: The data collection module is used to collect basic information about the current order when the current user enters the instant messaging chat interface. The basic information includes at least the order identification information of the current order. The acquisition module is used to acquire, when it is determined from the order identification information that the current order has historical dialogue information, the target text message, the sending client, the receiving client and the role type of the target text message in the historical dialogue information, wherein the target text message is the last text message in the historical dialogue information; The rewrite feedback module is used to generate a rewritten text message based on the role type of the sending client, and send the rewritten text message to the sending client so that the sending client replaces the target text message with the rewritten text message, and / or generate a feedback text message of the target text message based on the role type of the receiving client and recommend it to the receiving client.

[0007] Fourthly, this application provides a ride-hailing driver-passenger negotiation assistance device, which includes: The first receiving and display module is used to receive and display candidate rewritten text messages or candidate feedback text messages sent by the server when there is historical dialogue information in the instant messaging dialogue interface of the client, wherein the candidate rewritten text messages and the candidate feedback text messages are generated by the server based on the role type of the client. The replacement execution module is configured to respond to the current user's selection of a rewritten text message from the candidate rewritten text messages and replace the target text message with the rewritten text message, wherein the target text message is the last text message in the historical dialogue information; or, respond to the current user's reply operation based on the candidate feedback text message and perform a corresponding operation based on the reply operation. The second receiving and display module is used to receive and display a default recommended text message sent by the server when there is historical chat information in the instant messaging chat interface of the client. The default recommended text message is generated by the server based on the role type and order status information of the client. The selection and sending module is used to send the target recommended text message to the other party's client in response to the target recommended text message selected by the current user in the default recommended text message.

[0008] Fifthly, this application provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the ride-hailing driver-passenger negotiation assistance method as described in the first aspect above.

[0009] In a sixth aspect, this application provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the ride-hailing driver-passenger negotiation assistance method as described in the first aspect above.

[0010] In a seventh aspect, this application provides a chip including a processor and a communication interface, the communication interface being coupled to the processor, the processor being used to run programs or instructions to implement the ride-hailing driver-passenger negotiation assistance method as described in the first aspect.

[0011] Eighthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the ride-hailing driver-passenger negotiation assistance method as described in the first aspect above.

[0012] The above-described one or more technical solutions in the embodiments of this application have at least one of the following technical effects: 1. Increased negotiation success rate: By recommending AI-assisted negotiation scripts, the success rate of driver-passenger negotiations is expected to increase by 30%-40%, directly reducing the order cancellation rate caused by failed negotiations.

[0013] 2. Significantly improved communication efficiency: The one-click sending of AI-recommended scripts and the message rewriting and replacement function reduce the time users spend brainstorming and editing scripts from an average of 3-5 minutes to 3-5 seconds, improving communication efficiency by approximately 60 times.

[0014] 3. Controllable AI call costs: By reducing redundant LLM calls by 30%-50% through a multi-layer caching mechanism, the cost per call is approximately RMB 0.003-0.005. In scenarios with millions of daily dialogues, the monthly AI call cost is kept within an acceptable range.

[0015] 4. Ensuring User Experience Continuity: The streaming generation interruption-restart mechanism ensures that recommended content is synchronized with the latest conversation status in real time, and the cache status recovery mechanism ensures a seamless experience in page switching scenarios, with user satisfaction expected to increase by more than 25%.

[0016] 5. Cross-platform compatibility and scalability: The system supports operation on multiple platforms such as apps and mini-programs, and achieves a smooth transition between new and old versions through version compatibility strategies. The architecture design supports future expansion to more negotiation scenarios and AI capabilities.

[0017] Additional aspects and advantages of this application will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of this application. Attached Figure Description

[0018] The above and / or additional aspects and advantages of this application will become apparent and readily understood from the description of the embodiments taken in conjunction with the following drawings, in which: Figure 1 This is one of the flowcharts illustrating the ride-hailing driver-passenger negotiation assistance method provided in the embodiments of this application; Figure 2 This is a second schematic flowchart of the ride-hailing driver-passenger negotiation assistance method provided in the embodiments of this application; Figure 3 This is the third flowchart illustrating the ride-hailing driver-passenger negotiation assistance method provided in this application embodiment; Figure 4 This is the fourth flowchart of the ride-hailing driver-passenger negotiation assistance method provided in the embodiments of this application; Figure 5 This is the fifth flowchart illustrating the ride-hailing driver-passenger negotiation assistance method provided in the embodiments of this application; Figure 6 This is one of the structural schematic diagrams of the ride-hailing driver-passenger negotiation assistance device provided in the embodiments of this application; Figure 7 This is the second schematic diagram of the structure of the ride-hailing driver-passenger negotiation assistance device provided in the embodiments of this application; Figure 8 This is a schematic diagram of the structure of the electronic device provided in the embodiments of this application. Detailed Implementation

[0019] 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.

[0020] 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.

[0021] The following description, in conjunction with the accompanying drawings, details the ride-hailing driver-passenger negotiation assistance method, device, electronic device, and storage medium provided in this application through specific embodiments and application scenarios.

[0022] Among them, the ride-hailing driver-passenger negotiation assistance method can be applied to the terminal, and can be executed by the hardware or software in the terminal.

[0023] The ride-hailing driver-passenger negotiation assistance method provided in this application embodiment can be executed by an electronic device or a functional module or entity in an electronic device that can implement the ride-hailing driver-passenger negotiation assistance method. The electronic devices mentioned in this application embodiment include, but are not limited to, mobile phones, tablets, computers, cameras, and wearable devices. The ride-hailing driver-passenger negotiation assistance method provided in this application embodiment will be described below using an electronic device as the execution subject as an example.

[0024] like Figure 1 As shown in the embodiment of this application, the ride-hailing driver-passenger negotiation assistance method is applied to a server and includes the following steps: S110, Collect basic information of the current order when the current user enters the instant messaging chat interface, the basic information including at least the order identifier information of the current order; S120, when it is determined that the current order has historical dialogue information based on the order identification information, the target text message, the sending client, the receiving client and the role type of the target text message in the historical dialogue information are obtained, and the target text message is the last text message in the historical dialogue information; S130, a rewritten text message is generated based on the role type of the sending client, and the rewritten text message is sent to the sending client so that the sending client replaces the target text message with the rewritten text message, and / or, a feedback text message of the target text message is generated based on the role type of the receiving client and recommended to the receiving client.

[0025] The ride-hailing driver-passenger negotiation assistance method provided in this application collects basic information about the current order when it detects that a user has entered an IM (Instant Messaging) chat interface. This basic information includes at least the order identifier information of the current order. When it is determined that historical chat information exists in the instant messaging chat interface based on the order identifier information, the last text message in the historical chat information is obtained as the target text message. Simultaneously, the sending client, receiving client, and role type of the target text message are determined. On the sending client side, a rewritten text message is generated based on the sending client's role type and sent to the sending client, so that the sending client replaces the target text message with the rewritten text message, thereby realizing the IM chat interface... The solution rewrites the last text message sent in the context of the conversation. Simultaneously, on the receiving client side, it generates a feedback text message based on the receiving client's role type and recommends it to that client, thereby guiding the receiving client to make corresponding decisions. This solution makes communication between drivers and passengers more appropriate and efficient by intelligently rewriting the target text message and / or intelligently generating feedback text messages for the target text message. It avoids problems such as negotiation breakdowns due to inappropriate wording or poor communication methods. It is expected to reduce the negotiation failure rate from 25% to below 15%, reduce order cancellations due to negotiation failures by about 40%, improve communication efficiency and user experience, and increase the platform's order success rate.

[0026] The current user includes both the vehicle owner and the passenger, and the role type refers to whether the current user is the vehicle owner or the passenger. The sending client can be the vehicle owner's app, the passenger's app, or a passenger mini-program. The receiving client can be the passenger's app, passenger mini-program, or vehicle owner's app. In other words, when the sending client is the vehicle owner's app, the receiving client is the passenger's app or passenger mini-program; similarly, when the sending client is the passenger's app or passenger mini-program, the receiving client is the vehicle owner's app. Basic information includes order identifier information, order status information, the current user's role type, and the client version number.

[0027] It is worth noting that the rewritten text messages or feedback text messages generated by this application based on the current user's role type (i.e., the role type of the sending client or the role type of the receiving client) are more in line with the wording and wording of the actual driver-passenger communication scenario, which can further improve communication efficiency and user experience.

[0028] In some embodiments, such as Figure 2 As shown, step S130 includes: S131, the text content of the target text message, the historical dialogue information, and the role type are encapsulated into a rewrite request, and the rewrite request is sent to the large language model. The rewrite request is used to call the large language model to generate candidate rewritten text messages.

[0029] S132, receive the candidate rewritten text message generated by the large language model and push the candidate rewritten text message to the sending client, so that the sending client responds to the current user's selection in the candidate rewritten text message, replaces the target text message with the rewritten text message, and clears other candidate rewritten text messages, wherein the rewritten text message is the text message selected by the current user in the candidate rewritten text message, and the other candidate rewritten text messages are candidate rewritten text messages other than the rewritten text message.

[0030] In this embodiment, the text content of the target text message and the context in the historical dialogue information are encapsulated into a rewriting request. A large language model is called to intelligently generate candidate rewritten text messages, and the candidate rewritten text messages are pushed to the sending client. The sending client responds to the current user's selection, replaces the rewritten text message selected by the current user with the target text message, and clears other candidate rewritten text messages after the rewriting is completed. This achieves intelligent rewriting of the target text message and avoids the embarrassing situation where the current user finds that the wording is inappropriate after sending the message and is unable to correct it.

[0031] Among them, the large language model is LLM (Large Language Model).

[0032] In some embodiments, step S131 includes: Extract the tag information of the target text message, the tag information being used to indicate whether the target text message is text information generated by the large language model.

[0033] The text content of the target text message, the historical dialogue information, the role type, and the tag information are encapsulated into a rewrite request.

[0034] In this embodiment, when requesting the large language model to generate candidate rewritten text messages, a tag is added to indicate whether the target text message is text information generated by the large language model. The addition of this tag information can avoid repeated rewriting or excessive modification of the target text message, and can also record the current user's preference for AI rewriting, thereby optimizing the subsequent AI rewriting process.

[0035] For example, if the target text message adopts AI-generated content, select the "tone adjustment" rewriting option; if the target text message does not adopt AI-generated content, select the "professional enhancement" rewriting option.

[0036] In some embodiments, when encapsulating the rewrite request, all information in the basic information (such as order status information, client version number, etc., not included in the aforementioned embodiments) can also be encapsulated into a rewrite request, thereby making the generated candidate rewrite text messages more in line with the wording and wording of the actual driver-passenger communication scenario, and improving the accuracy of the wording. In some embodiments, the client version number can allow the large language model to determine whether to generate candidate feedback text messages (e.g., older versions do not support the rewrite function).

[0037] In some embodiments, while performing steps S131 and S132 as described above, the method further includes: A first prompt message is generated and displayed within a first preset range of the target text message. The first prompt message is used to indicate that the target text message is being rewritten.

[0038] This embodiment can prompt the receiving client that the current user's target text message is being modified and that they can wait to obtain the true intent of the sending client. The first preset range can be below, above, in front of, or behind the target text message. Preferably, the first prompt is positioned below the target text message. For example, the first prompt can be "polishing," "rewriting," or "editing."

[0039] In some embodiments, after step S132, the method further includes: A second prompt message is generated and displayed within a second preset range of the rewritten text message. The second prompt message is used to indicate that the target text message has been rewritten.

[0040] This embodiment can provide a notification that the rewriting of the target text message is complete. The second preset range can be below, above, in front of, or behind the target text message. Preferably, the second notification information is positioned below the target text message. For example, the first notification information can be "polished," "rewriting in progress," or "edited."

[0041] In some embodiments, while performing step S132, the method further includes: Generate a first virtual button for each of the candidate rewritten text messages, and configure the first virtual button to be displayed within a third preset range of the corresponding candidate rewritten text message. The first virtual button is used to respond to the current user's selection of the candidate rewritten text message.

[0042] This embodiment enables client users to select candidate rewritten text messages. The third preset range can be located below, above, in front of, behind, or superimposed on the candidate rewritten text messages. Preferably, the first virtual button is positioned behind the candidate rewritten text messages. For example, the first virtual button can be displayed as "Replace" or "Select".

[0043] It should be noted that a first virtual button can be displayed within a third preset range for each candidate rewritten text message, and the user's selection of this first virtual button means selecting the corresponding candidate rewritten text message; alternatively, a first virtual button can be displayed within a third preset range for all candidate rewritten text messages, and the user selects the corresponding candidate rewritten text message by selecting its sequence number. For example, there can be 1 to 5 candidate rewritten text messages. Preferably, there are 2 candidate rewritten text messages. In some embodiments, the number of characters in the candidate rewritten text messages meets a preset character limit. In some embodiments, the number of characters in the candidate rewritten text messages is less than or equal to 30 characters.

[0044] In some embodiments, such as Figure 3 As shown, the basic information also includes order status information, and step S130 further includes: S133, the historical dialogue information, the role type, and the order status information are encapsulated into a prompt request, and the prompt request is sent to the large language model. The prompt request is used to call the large language model to generate candidate feedback text messages.

[0045] S134, receive the candidate feedback text message generated by the large language model and push the candidate feedback text message to the receiving client, so that the receiving client responds to the current user's reply operation based on the candidate feedback text message.

[0046] In this embodiment, when generating candidate feedback text messages by calling the large language model on the receiving client side, not only historical dialogue information and the role type of the receiving client are used, but also order status information is used. This makes the generated candidate feedback text messages more in line with the actual scenario and avoids abrupt candidate feedback text messages.

[0047] It's worth noting that the candidate feedback text messages include candidate reply text messages and recommendation suggestion text messages. The current user on the receiving client can select a response based on the candidate reply text messages, while the current user on the receiving client can simply view the recommendation suggestion text messages without replying. The order status information includes various real-time states such as pending order, accepted order, pending departure, departed, and completed.

[0048] In some embodiments, when encapsulating the prompt request, all information in the basic information (such as the client version number not included above) can also be encapsulated into a rewrite request, thereby making the generated candidate feedback text message more in line with the wording and wording of the actual driver-passenger communication scenario, and improving the accuracy of the wording. In some embodiments, the client version number can allow the large language model to decide whether to generate candidate feedback text messages (e.g., older versions do not support the smart reply function).

[0049] In some embodiments, while performing steps S133 and S134, the method further includes: A third prompt message is generated and configured to be displayed within a fourth preset range of the instant messaging dialog interface. The third prompt message is used to indicate that the candidate feedback text message is being generated.

[0050] This embodiment can indicate to the receiving client that the current user candidate feedback text message is being generated, allowing the user to wait, edit the reply text message, or execute the corresponding strategy. The fourth preset range can be an input box or dialog box in the instant messaging interface. Preferably, the third prompt information is positioned above or below the interface between the input box and the dialog box in the instant messaging interface. For example, the third prompt information may be an animation such as "Thinking," "Generating," "Please wait," or "Inputting."

[0051] In some embodiments, when the candidate feedback text message is the candidate reply text message, step S134 includes: A second virtual button is generated for each of the candidate feedback text messages, and the second virtual button is configured to be displayed within a fifth preset range of the corresponding candidate feedback text message. The second virtual button is used to respond to the current selection of the candidate feedback text message.

[0052] This embodiment enables the receiving client user to select candidate reply text messages. The fifth preset range can be located below, above, in front of, behind, or superimposed on the candidate reply text messages. Preferably, the second virtual button is configured behind the candidate reply text messages. For example, the second virtual button can be displayed as "Send" or "Select".

[0053] It should be noted that a second virtual button can be displayed within a fifth preset range for each candidate reply text message, and the user's selection of this second virtual button means selecting the corresponding candidate reply text message; alternatively, a second virtual button can be displayed within a fifth preset range for all candidate reply text messages, and the user selects the corresponding candidate reply text message by selecting its sequence number. For example, there can be 1 to 5 candidate reply text messages. Preferably, there are 2 candidate reply text messages. In some embodiments, the number of characters in the candidate reply text messages meets a preset character limit. In some embodiments, the number of characters in the candidate reply text messages is less than or equal to 30 characters.

[0054] In some embodiments, while performing step S134, the method further includes: The size of the first floating window is adjusted according to all the candidate feedback text messages so that all the candidate feedback text messages are displayed within the first floating window.

[0055] This embodiment enables the complete display of all candidate feedback text messages in the IM conversation interface, improving the readability of the current candidate feedback text messages on the receiving client side.

[0056] In some embodiments, the large language model generates the candidate feedback text messages in a streaming manner. The generation time of the candidate feedback text messages by the large language model is on the order of seconds. Further, the generation time is 1.8 to 2.5 seconds.

[0057] It should be noted that the streaming method refers to displaying one character on the IM chat interface as soon as the large language model generates one character. For example, the generation of the first character takes 1.8 seconds, and the generation of the last character (that is, the entire candidate feedback text is generated) takes 2.5 seconds. This can shorten the original process of devising a script to a second-level operation. It solves the contradiction between the generation delay of the large language model and the real-time requirements of instant communication, and greatly improves communication efficiency.

[0058] The above steps S110~S130 and their specific implementation methods can be implemented by the following algorithm code, specifically: Input: ContextVector = {OrderStatus, UserRole, DialogHistory,MessageType} Output: StrategyType, ContentConfig Step 1: Determine the dialogue history IF DialogHistory.count == 0: StrategyType = "DEFAULT_RECOMMEND" ContentConfig = getDefaultConfig(OrderStatus, UserRole) ELSE: goto Step 2 Step 2: Analyze the last message LastMessage = DialogHistory.getLastTextMessage(type=1) IF LastMessage.sender == "COUNTERPART": StrategyType = "REPLY_RECOMMEND" goto Step 3 ELIF LastMessage.sender == "SELF": StrategyType = "REWRITE_RECOMMEND" goto Step 4 Step 3: Generate recommended reply / suggestion Features = extractFeatures(DialogHistory, OrderStatus, UserRole) IF Features.needSuggestion: ContentConfig = {type: "SUGGESTION", maxCount: 1, maxLength: 40} ELSE: ContentConfig = {type: "REPLY", maxCount: 3, maxLength: 40} Step 4: Generate recommended rewrite IsAIAdopted = LastMessage.hasTag("ai_adopted") ContentConfig = {type: "REWRITE", maxCount: 2, maxLength: 30,aiAdopted: IsAIAdopted} Returns: {StrategyType, ContentConfig} In the algorithm code above, the meanings of the different dimensions of information are as follows: Order Status Dimension: The order lifecycle is divided into 6 status stages (Status 10 - Pending Order, Status 20 - Order Accepted, Status 30 - Pending Departure / Departure, Status 40 / 50 / 60 - During Trip and Subsequent Status), with each status corresponding to different negotiation focuses and communication styles.

[0059] User Role Dimension: Distinguishing between Driver and Passenger. The two roles have fundamentally different negotiation stances, ways of expressing themselves, and concerns—Drivers are more concerned with time efficiency and cost sharing, while passengers are more concerned with service experience and price reasonableness.

[0060] Dialogue History Dimension: Analyzes historical dialogue records under the current order ID dimension to extract features such as dialogue rounds, negotiation topics (cost / time / route), and sentiment (positive / neutral / negative).

[0061] Message Type Dimension (MessageType): Analyze the type of the last message (type=1 text message filtering system message), the sender (self / other party), and whether it is AI-recommended content (aitype tag).

[0062] In some embodiments, such as Figure 4 As shown, the basic information also includes order status information, and the method further includes: S140, when it is determined from the order identification information that there is no historical dialogue information in the instant messaging dialogue interface, a default recommended text message is obtained based on the current user's role type and the order status information.

[0063] S150, the default recommended text message is pushed to the current user's client, so that the client responds to the target recommended text message selected by the current user in the default recommended text message and sends the target recommended text message to the other party's client; wherein, the client includes a sending client and a receiving client.

[0064] In this embodiment, when it is determined that there is no historical dialogue information in the IM dialogue interface, the system will generate different default recommended text messages based on the current user's role type and order status information, and push the default recommended text messages to the client. Thus, the client responds to the current user's selection and determines the target recommended text message to be sent to the other party's client, which can realize efficient and smooth communication between the driver and passenger during the initial dialogue.

[0065] It is worth noting that the client in this embodiment includes a sending client and a receiving client. That is, the client can be a car owner's APP client, a passenger's APP client, or a passenger's mini-program client.

[0066] In some embodiments, when obtaining the default recommended text message, the default recommended text message can also be obtained together with all information in the basic information (such as the client version number, which was not included in the aforementioned embodiments). This makes the generated default recommended text message more in line with the wording and wording of the actual driver-passenger communication scenario, thereby improving the accuracy of the wording. In some embodiments, the client version number can determine whether to obtain the default recommended text message (e.g., older versions do not support the function of intelligently generating default recommended text messages).

[0067] In some embodiments, obtaining the default recommended text message may involve encapsulating some or all of the basic information into a request to request the large language model to generate the default recommended text message.

[0068] In some embodiments, while performing step S150, the method further includes: A fourth prompt message is generated and configured within a sixth preset range of the second floating window. The second floating window is used to display the default recommended text message, and the fourth prompt message is used to indicate the title information of the second floating window.

[0069] This embodiment enables the complete display of default recommended text messages in the second floating window and the prompting of system-generated default recommended text messages. The sixth preset range can be above or below the second floating window, which can be located in the input box or dialog box of the IM chat interface. In some embodiments, the fourth prompt message is configured above the second floating window, which is located at the boundary between the input box and the dialog box of the IM chat interface. For example, the fourth prompt message is "Use AI for negotiation, making communication smoother~".

[0070] In some embodiments, while performing step S150, the method further includes: A third virtual button is generated for each of the default recommended text messages, and the third virtual button is configured to be displayed within a seventh preset range of the corresponding default recommended text message. The third virtual button is used to respond to the current user's selection of the default recommended text message.

[0071] This embodiment enables the selection of a default recommended text message. The seventh preset range can be located below, above, in front of, behind, or superimposed on the candidate reply text message of the default recommended text message. Preferably, the third virtual button is configured behind the default recommended text message. For example, the third virtual button can be displayed as "Send" or "Select".

[0072] It should be noted that a third virtual button can be configured to be displayed within the seventh preset range of each default recommended text message, and the user's selection of this third virtual button means selecting the corresponding default recommended text message; alternatively, a third virtual button can be displayed within the seventh preset range of all default recommended text messages, and the user selects the corresponding default recommended text message by selecting its sequence number. For example, there can be 1 to 5 default recommended text messages. Preferably, there are 2 default recommended text messages. In some embodiments, the number of characters in the default recommended text messages meets a preset character limit. In some embodiments, the number of characters in the default recommended text messages is less than or equal to 30 characters.

[0073] Through the above embodiments, at least the following technical effects can be achieved: 1. Significantly reduce order cancellation rate due to negotiation failures: By using AI-generated professional negotiation scripts to assist drivers and passengers in communication, the negotiation failure rate is expected to be reduced from 25% to below 15%, reducing order cancellations due to negotiation failures by approximately 40%, directly improving the platform's order completion rate. 2. Significantly improve driver-passenger communication efficiency and negotiation experience: The response time for the first character of the AI-recommended reply is approximately 1.8 seconds, and the complete generation time is approximately 2.5 seconds. Users can send recommended scripts with one click or replace and rewrite content with one click, shortening the script conception process that originally required several minutes to a second-level operation, improving communication efficiency by more than 80%. 3. Achieve differentiated negotiation strategy coverage throughout the entire order lifecycle: The system supports differentiated recommendation strategies throughout the entire process from order status 10 (pending acceptance) to status 60 (completed), ensuring that users entering IM chat at any stage can receive negotiation assistance matching the current scenario. 4. Intelligent message rewriting capability expands the space for communication optimization: The unique sent message rewriting and replacement mechanism allows users to still have the opportunity to optimize their expression after sending a message. Combined with the cross-platform "edited" mark, it improves the quality of communication while ensuring information transparency.

[0074] In some embodiments, during the execution of the aforementioned steps S130-S150, the method further includes: The instant messaging interface is continuously monitored, and when a new target text message is detected, the process of generating the rewritten text message, generating the feedback text message, or obtaining the default recommended text message is interrupted, the currently displayed generated content is cleared, and the process returns to step S120; wherein, the generated content includes at least one of the rewritten text message, feedback text message, or default recommended text message generated during the generation process.

[0075] Through this embodiment, when a new message is generated in a conversation, the system can immediately interrupt the current AI generation and start a new round of recommendations, ensuring that the recommended content is always synchronized with the latest conversation status and avoiding a decline in user experience caused by recommendation lag.

[0076] Understandably, in IM conversation scenarios, when a new text message is detected, the current LLM call is immediately canceled, the displayed content is cleared, the state machine is reset to IDLE, and the context is reconstructed based on the new message, a strategy is selected, and a new round of generation is triggered. In IM rewriting scenarios, if a new message is generated during the recommendation rewriting generation process, the system not only interrupts the rewriting generation but also automatically switches the recommendation mode (from "rewrite" to "recommend reply," etc.) based on the sender of the new message.

[0077] In some embodiments, the streaming content generation and interruption-restart control mechanism for instant messaging scenarios can be expressed as follows: Streaming generation state machine: Defines a four-state finite state machine (IDLE→LOADING→STREAMING→COMPLETED) to manage the entire lifecycle of streaming content generation, including loading animation display, sequential content streaming display, button interaction state switching, interruption clearing, and other complete state transitions.

[0078] State Machine Definition: StreamingFSM State set: {IDLE, LOADING, STREAMING, COMPLETED} Conversion rules: IDLE → LOADING: Triggering condition: Receiving an AI recommendation request Actions: Display a loading animation ("Thinking..."), and initiate an LLM streaming API call. LOADING → STREAMING: Triggering condition: Receiving the first content fragment returned by the LLM (the first character is delayed by approximately 1.8 seconds). Action: Hide the loading animation and begin displaying recommended content item by item. STREAMING → STREAMING: Trigger condition: Receiving subsequent content fragments Action: Display new content; if it's a recommended reply, highlight the corresponding "Send" button. STREAMING → COMPLETED: Triggering condition: LLM returns the end marker. Action: Adaptively adjust the floating window height; all buttons become interactive. {LOADING, STREAMING} → IDLE: Triggering condition: A new message sending event (interrupt signal) is detected. Action: Immediately terminate the current LLM call, clear the displayed content, and re-enter the scene for judgment. {LOADING, STREAMING, COMPLETED} → IDLE: Triggering condition: The user clicks the close button or pulls up the keyboard. Actions: Hide the floating window / rewrite the prompt, and switch the AI ​​negotiation assistant button to an unselected state. In some embodiments, the method further includes: When it is detected that the current user has exited the instant messaging chat interface, the current AI recommendation state is not saved.

[0079] When the current user re-enters the instant messaging chat interface, a target AI type message is obtained. The target AI type message is the last text message of type AI sent by the current user in the historical chat information.

[0080] Extract the target message content and target message identifier of the target AI type message. Use the target message content and target message identifier to query the target cache. If there is valid AI recommendation content in the target cache, push the AI ​​recommendation content to the current user's client. If there is no valid AI recommendation content in the target cache, encapsulate the target message content and target message identifier into a call request and send it to the large language model. The call request is used to request the large language model to regenerate valid AI recommendation content and push it to the current user's client.

[0081] This embodiment proposes an AI recommendation result caching layering and request deduplication mechanism for high-frequency IM chat interface switching scenarios. This avoids repeated AI calls caused by users frequently entering and exiting the IM chat interface. It can achieve cost control of about 0.003-0.005 yuan per call for large language models, effectively reducing the number of repeated calls and computational costs of large language models, and achieving cost control in large-scale user scenarios.

[0082] For example: After a passenger exits and re-enters the IM interface, the system retrieves the latest message (the driver replies "Okay, 8:30 no problem"), determines it's not an AI-type message and was sent by the other party → needs to request a recommendation → the service query cache misses → requests the AI ​​interface to generate a new recommendation "Thank you for understanding! See you tomorrow morning at 8:30~" → updates the cache and displays it.

[0083] Specifically, the cache key design and storage strategy in this embodiment are as follows: A composite cache key scheme based on content + messageId (content + identifier, corresponding to the aforementioned target message content and target message identifier) ​​is designed at the reach service layer, supporting cache storage with TTL and validity judgment; Page re-entry request decision logic: When a user re-enters the page, the system determines whether a recommendation request is needed by judging the type, sender, and AI tag of the latest text message; if needed, the cache is queried first, and if the cache is hit, it is returned directly; otherwise, the AI ​​interface is requested and the cache is updated; Cost control effect: Each AI call consumes approximately 600 tokens of input and 20 tokens of output, with a single cost of approximately 0.003-0.005 yuan. The caching mechanism is expected to reduce redundant calls by 30%-50%.

[0084] In some embodiments, the method further includes: When an adoption event is detected that the current user has adopted the event, the adoption information of the adoption event is recorded. The adoption event includes at least one of the rewritten text message, the reply text message, and the default recommended text message being adopted by the current user. The adoption information includes at least one of the adoption content, the type of adoption content, the order of adoption content, the order status of adoption content, and the role type of the adopting user.

[0085] The adopted information is used to iteratively train the large language model.

[0086] In this embodiment, by tracking and feedback looping adoption behavior, adoption events are recorded while the adoption information is used to iteratively process the large language model, achieving the effect of supporting offline analysis and model iterative optimization. Specifically, before using the adoption information to iteratively train the large language model, the adoption information can be synchronously written into a base table and labeled to improve model training efficiency.

[0087] In some embodiments, the basic information may also include the version information of the current user's client; If the client's version information is a preset version, the client's instant messaging interface is controlled to display the original text message; if the client's version information is not a preset version, the client's instant messaging interface is controlled to display an AI text message. The original text message is the text message edited by the current user, and the AI ​​text message is a text message generated based on a large language model, including at least one of rewritten text messages, feedback text messages, and default recommended text messages.

[0088] In this embodiment, compatibility between new and old versions of the vehicle owner APP client, passenger APP client, and passenger mini-program client can be achieved. Through version compatibility logic (the new version displays AI text messages, and the old version displays original text messages), a seamless experience for users of different versions is ensured.

[0089] like Figure 5 As shown in the embodiments of this application, the ride-hailing driver-passenger negotiation assistance method is applied to the client and includes the following steps: S210, when historical conversation information exists in the instant messaging interface of the client, receive and display candidate rewrite text messages or candidate feedback text messages sent by the server, wherein the candidate rewrite text messages and candidate feedback text messages are generated by the server based on the role type of the client.

[0090] S220, in response to the current user selecting a rewritten text message from the candidate rewritten text messages, replace the target text message with the rewritten text message, wherein the target text message is the last text message in the historical dialogue information. Alternatively S230, responding to the current user's reply operation based on the candidate feedback text message, and performing corresponding operations based on the reply operation.

[0091] S240, when historical chat information exists in the instant messaging chat interface of the client, a default recommended text message sent by the server is received and displayed, wherein the default recommended text message is generated by the server based on the role type and order status information of the client.

[0092] S250, in response to the target recommended text message selected by the current user in the default recommended text message, the target recommended text message is sent to the other party's client.

[0093] The ride-hailing driver-passenger negotiation assistance method provided in this application embodiment, when historical dialogue information exists in the client's instant messaging interface, if the client is the sending client of the target text message, then the sending client receives and displays the candidate rewritten text message sent by the server. Next, the sending client responds to the current user's (i.e., the sender's) selection, replacing the target text message in the IM dialogue interface with the rewritten text message selected by the current user, thereby rewriting the last text message in the sent IM dialogue interface. Simultaneously, if the client is the receiving client of the target text message, then the receiving client receives and... The server displays candidate feedback text messages, and then the receiving client responds to the current user's (i.e., the recipient's) reply by performing corresponding operations, thereby achieving intelligent reply from the receiving client. This solution makes communication between drivers and passengers more appropriate and efficient by intelligently rewriting and / or intelligently generating feedback text messages for the target text messages, avoiding the problem of negotiation breakdowns due to inappropriate wording or poor communication methods. It is expected to reduce the negotiation failure rate from 25% to below 15%, reduce order cancellations due to negotiation failures by about 40%, improve communication efficiency and user experience, and increase the platform's order success rate.

[0094] Understandably, the current users include both vehicle owners and passengers, and the role type refers to whether the current user is a vehicle owner or a passenger. The sending client can be a vehicle owner's app client, a passenger's app client, or a passenger's mini-program client, and the receiving client can be a passenger's app client, a passenger's mini-program client, or a vehicle owner's app client. It's understood that when the sending client is a vehicle owner's app client, the receiving client is a passenger's app client or a passenger's mini-program client, and similarly, when the sending client is a passenger's app client or a passenger's mini-program client, the receiving client is a vehicle owner's app client. Basic information includes order identification information, order status information, the current user's role type, and the client version number.

[0095] Furthermore, when there is no historical conversation information in the client's instant messaging interface, the client (including the driver's client and the passenger's client) will receive and display the default recommended text message sent by the server. Then, the client (driver's client or passenger's client) responds to the current user's selection and sends the target recommended text message to the other client (corresponding passenger's client or driver's client). This solution can achieve efficient and smooth communication between the driver and passenger during the initial conversation.

[0096] Understandably, the default recommended text messages for the driver's client and the passenger's client can be the same or different. Preferably, the default recommended text messages for the driver's client and the passenger's client are different. The passenger client includes a passenger APP client or a passenger mini-program client, and the driver's client includes a driver APP client.

[0097] It should be noted that the candidate rewritten text message and the candidate feedback text message are generated by the server based on the role type of the client, including: the candidate rewritten text message is generated by the server based on the role type of the sending client, where the sending client is the client that sends the target text message; the candidate feedback text message is generated by the server based on the role type of the receiving client, where the receiving client is the client that receives the target text message.

[0098] In some embodiments, the candidate rewritten text message may also be generated by the server based on the role type of the sending client and basic information.

[0099] In some embodiments, the candidate feedback text message may also be generated by the server based on the role type and basic information of the receiving client.

[0100] The basic information includes at least one of the following: order identification information, order status information, current user role type, and client version number.

[0101] In some embodiments, the candidate feedback text message includes a candidate reply text message and a recommendation suggestion text message. Step S230 includes: If the candidate feedback text message is the candidate reply text message, in response to the reply text message selected by the current user in the candidate reply text message, the reply text message is sent to the other party's client.

[0102] If the candidate feedback text message is the recommendation suggestion text message, then the recommendation suggestion text message will be displayed in the instant messaging dialog interface.

[0103] In this embodiment, the candidate feedback text message includes candidate reply text messages and recommendation suggestion text messages. The current user of the receiving client can select and reply to the target text message based on the candidate reply text message, while the current user of the receiving client can only view the recommendation suggestion text message without replying. This solution can realize personalized operations for different scenarios and situations.

[0104] In some embodiments, the method further includes: Before replacing the target text message with the rewritten text message, a first prompt message is displayed within a first preset range of the target text message. The first prompt message is used to indicate that the target text message is being rewritten. After the target text message is replaced with the rewritten text message, a second prompt message is displayed within a second preset range of the rewritten text message. The second prompt message is used to indicate that the target text message has been rewritten. While receiving and displaying candidate rewritten text messages sent by the server, a first virtual button is displayed within a third preset range of each candidate rewritten text message according to the server's configuration. The first virtual button is used to respond to the current user's selection of the candidate rewritten text message.

[0105] This embodiment enables prompts for the sending client in different states, further improving the transparency and completeness of the intelligent rewriting function. The first virtual button allows the sending client user to select candidate rewritten text messages, enhancing the convenience of the rewriting function.

[0106] In some embodiments, the method further includes: Before receiving and displaying the candidate feedback text message sent by the server, the third prompt information is displayed within a fourth preset range of the instant messaging dialog interface. The third prompt information is used to indicate that the candidate feedback text message is being generated. While receiving and displaying candidate feedback text messages sent by the server, a second virtual button is displayed within a fifth preset range corresponding to each candidate feedback text message according to the server's configuration. The second virtual button is used to respond to the current user's selection of the candidate feedback text message. The process of receiving and displaying the candidate feedback text message sent by the server includes: displaying the candidate feedback text message in a first floating window according to the server's adjustment.

[0107] This embodiment enables the receiving client to receive prompts in different states, further improving the transparency and completeness of the intelligent reply function. The second virtual button and the first floating window allow the receiving client user to select and visualize candidate feedback text messages, enhancing the convenience of the intelligent reply function.

[0108] In some embodiments, the method further includes: While receiving and displaying the default recommended text message sent by the server, a fourth prompt message is displayed within a sixth preset range of the second floating window according to the server's configuration, and a third virtual button is displayed within a seventh preset range corresponding to each of the default recommended text messages. The fourth prompt message is used to indicate the title information of the second floating window, the second floating window is used to display the default recommended text message, and the third virtual button is used to respond to the current user's selection of the default recommended text message.

[0109] In this embodiment, the second floating window and the fourth prompt information can improve the visualization effect of the default recommended text message, and the third virtual button can facilitate the selection of the default recommended text message.

[0110] In some embodiments, the text message in this application may be represented by "Type=1". It should be understood that the sequence numbers in the foregoing embodiments do not constitute a limitation on the order of the steps, but are only for explaining the relevant technical processes. For example, step S131 may occur simultaneously with step S133, and steps S130 and S140 may occur concurrently.

[0111] It should be noted that the corresponding technical features of the ride-hailing driver-passenger negotiation assistance method applied to the client and the ride-hailing driver-passenger negotiation assistance method applied to the server provided in this application have the same technical effects. To avoid duplication, they will not be described again here.

[0112] To better understand this application, the following description, in conjunction with specific scenarios, illustrates the aforementioned implementation methods and their effects: Example 1: Scenario of car owners negotiating highway tolls Scene description: Mr. Zhang, the driver, received a rideshare order from Shanghai to Hangzhou. The order required taking the highway, incurring a toll of approximately 80 yuan. The passenger did not explicitly state whether they were willing to share the toll when placing the order. Mr. Zhang attempted to negotiate the toll-sharing arrangement with the passenger via instant messaging.

[0113] Implementation steps: Step 1: Context Awareness and Scene Determination When a car owner enters the IM chat interface, the system collects the following context information: OrderStatus=20 (order accepted), UserRole=Driver (car owner), DialogHistory=0 (no history of conversations), ClientVersion=6.96.5 (supports all functions). The system determines: no history of conversations → triggers the default recommendation display strategy.

[0114] Step 2: Default Recommendation Generation and Display Based on the combination OrderStatus=20+UserRole=Driver, the system retrieves the default recommended message after the driver accepts the order from the policy library: 1. "Hello, can we negotiate to share the highway toll?" 2. "Please confirm the departure time so I can make arrangements." 3. "The trip will take approximately X hours, and tolls will be charged for highway travel." 4. "Where exactly is your pick-up point?" The system displays this message in a floating window with the title "Use AI for negotiation, making communication smoother~".

[0115] Step 3: Car owner adopts recommendation The car owner clicks the "Send" button for the first recommendation, and the message is sent to the chat with one click. The system records the adoption event, including information such as content_type=3 (default recommendation), content_order=1 (first recommendation), and role=2 (car owner).

[0116] Step 4: AI generates recommended responses The passenger replied, "The highway toll is too expensive; I don't think it's reasonable." The system detected that the last message was sent by the other party (the passenger) and triggered the generation of recommended replies. The recommended replies are generated using LLM streaming (first character delayed by 1.6 seconds, complete generation in 2.3 seconds), containing 2 recommended replies and 1 recommendation suggestion.

[0117] Step 5: Message rewriting optimization After the car owner selects to send the first recommended reply, the system triggers a rewrite recommendation. AI generates two rewrite candidates, and the car owner selects the more convincing rewrite content and clicks "Replace." The system then replaces the original message with the rewritten content, and the passenger's message is updated and displays an "Edited" label.

[0118] The above embodiment 1 can achieve at least the following technical effects: It only takes 3 seconds for a car owner to go from entering the page to sending the first negotiation message (including selection and clicking); with the professional language recommended by AI, the probability of reaching an agreement on highway tolls increases by about 35%; the rewrite function makes car owners express themselves more professionally and persuasively, reducing communication friction caused by inappropriate wording.

[0119] Example 2: Passengers negotiate travel time adjustments Scene description: Ms. Li, a passenger, booked a rideshare order for 8:00 AM the following morning, but needed to change the departure time to 8:30 AM. The driver accepted the order, and the two had a brief greeting. Ms. Li then requested to discuss the time change with the driver via instant messaging.

[0120] Implementation steps: Step 1: Context Awareness and Scene Determination When a passenger enters the IM chat page, the system collects the following context information: OrderStatus=30 (pending departure - not yet departed), UserRole=Passenger, DialogHistory=2 (there was a simple greeting conversation), and LastMessage was sent by the driver as "Okay, see you tomorrow morning." The system determines that there is a history of conversations, and the last message was sent by the driver → triggering the generation of recommended replies / suggestions.

[0121] Step 2: AI-generated recommended responses The system encapsulates the dialogue history and order information into a prompt, which is then generated using an LLM (Large Language Model). The process is as follows: 1. A floating window displays a loading animation "Thinking..." (1.8 seconds); 2. Streaming output includes recommended reply 1 "Sorry, something came up, could we reschedule for 8:30?", recommended reply 2 "Could we postpone the departure by half an hour? Sorry for the inconvenience~", and a suggestion "You can use the 'Modify Time' function below to directly submit a time change request"; 3. Once all data is generated, the floating window height automatically adjusts.

[0122] Step 3: Passengers use the quick access functions in the function area After seeing the recommendations, passengers noticed the "Modify Time" button in the function operation area. Clicking it brought up a time selection component, where passengers selected 8:30 as the new departure time, and the system sent a time modification message.

[0123] Step 4: Streaming Interrupts and Restarts If the previous recommendation pop-up is still displayed while the passenger is sending a time change message, the system immediately interrupts and clears the pop-up content. Based on the new conversation state, the system determines not to trigger a new recommendation response.

[0124] Step 5: Caching and Reentrancy Handling After the passenger exits and re-enters the IM page, the system retrieves the latest message (the driver replies "Okay, 8:30 no problem"), determines it's not an AI-type message and was sent by the other party → requests a recommendation → the service query cache is not hit → requests the AI ​​interface to generate a new recommendation "Thank you for understanding! See you tomorrow morning at 8:30~" → updates the cache and displays it.

[0125] The above-described embodiment 2 can achieve at least the following technical effects: Passengers can express their time adjustment requests using AI-recommended phrases, making their wording more appropriate and polite; the exposed "Modify Time" function area shortens the operation path by 50%, eliminating the need for users to search through menus; the caching mechanism reduces the delay in obtaining recommended results when passengers re-enter the page to less than 200ms (when the cache is hit); and the entire negotiation process time is reduced from an average of 5 minutes to less than 2 minutes.

[0126] Based on the foregoing description, this application has at least the following technical advantages compared to the prior art: Traditional IM quick replies provide a fixed, preset list of quick phrases, the content of which does not change based on the conversation context. This application uses a large language model to understand the real-time conversation context and dynamically generates script recommendations that highly match the current negotiation scenario. It supports four different forms of recommended content (default recommendation, recommended suggestion, recommended reply, and recommended rewrite), covering the entire negotiation process. In particular, the recommended rewrite capability enables intelligent optimization and cross-platform replacement of sent messages, a capability completely lacking in traditional quick reply technology.

[0127] The general AI input method provides high-EQ communication suggestions for general scenarios, without optimization for specific industries or business scenarios. This invention deeply integrates knowledge from ride-hailing and carpooling businesses, generating targeted scripts based on multi-dimensional business characteristics such as order status (6 stages), user roles (driver / passenger), and negotiation topics (cost / time / route). The system integrates IM-specific technical capabilities such as a streaming generation interruption-restart mechanism, cross-platform message replacement synchronization, and multi-layer cache state recovery.

[0128] Existing chatbots typically act as participants in conversations, responding on behalf of humans. In this application, the AI ​​Agent acts as a "negotiation assistant" for the user, rather than a participant in the conversation, providing only recommendations and suggestions; the final sending decision is entirely controlled by the user. The system innovatively offers two unique auxiliary modes: "recommendation suggestions" (for reference only and not to be sent) and "recommendation rewriting" (optimizing sent content).

[0129] The ride-hailing driver-passenger negotiation assistance method provided in this application can be executed by a ride-hailing driver-passenger negotiation assistance device. This application uses the ride-hailing driver-passenger negotiation assistance device executing the method as an example to illustrate the ride-hailing driver-passenger negotiation assistance device provided in this application.

[0130] This application also provides a ride-hailing driver-passenger negotiation assistance device, such as... Figure 6 As shown, the ride-hailing driver-passenger negotiation assistance device includes: The data collection module 110 is used to collect basic information of the current order when the current user enters the instant messaging chat interface. The basic information includes at least the order identification information of the current order.

[0131] The acquisition module 120 is used to acquire, when it is determined from the order identification information that the current order has historical dialogue information, the target text message, the sending client, the receiving client and the role type of the target text message in the historical dialogue information, wherein the target text message is the last text message in the historical dialogue information.

[0132] The rewrite feedback module 130 is used to generate a rewritten text message based on the role type of the sending client, and send the rewritten text message to the sending client so that the sending client replaces the target text message with the rewritten text message, and / or generate a feedback text message of the target text message based on the role type of the receiving client and recommend it to the receiving client.

[0133] According to the ride-hailing driver-passenger negotiation assistance device provided in this application embodiment, when the current user enters the IM chat interface, basic information of the current order is collected. The basic information includes at least the order identifier information of the current order. When it is determined that there is historical chat information in the instant messaging chat interface based on the order identifier information, the last text message in the historical chat information is obtained as the target text message. At the same time, the sending client, receiving client, and role type of the target text message are determined. On the sending client side, rewritten text information is generated based on the role type of the sending client and sent to the sending client so that the sending client replaces the target text message with the rewritten text message, thereby realizing the rewriting of the last text message sent in the IM chat interface. At the same time, on the receiving client side, feedback text messages are generated based on the role type of the receiving client and recommended to the receiving client, thereby guiding the receiving client to make corresponding decisions. This solution makes the communication measures between the driver and passenger more appropriate and efficient by intelligently rewriting the target text message and / or intelligently generating feedback text messages for the target text message, avoiding the problem of negotiation breakdown due to inappropriate wording or poor communication methods, improving communication efficiency and user experience, and increasing the platform's order success rate.

[0134] The ride-hailing driver-passenger negotiation assistance device provided in this application embodiment can achieve... Figures 1 to 4 The various processes implemented in the method implementation examples will not be described again here to avoid repetition.

[0135] This application also provides a ride-hailing driver-passenger negotiation assistance device, such as... Figure 7 As shown, the ride-hailing driver-passenger negotiation assistance device includes: The first receiving and display module 210 is used to receive and display candidate rewritten text messages or candidate feedback text messages sent by the server when there is historical dialogue information in the instant messaging dialogue interface of the client, wherein the candidate rewritten text messages and candidate feedback text messages are generated by the server based on the role type of the client.

[0136] The replacement execution module 220 is configured to respond to the current user's selection of a rewritten text message from the candidate rewritten text messages and replace the target text message with the rewritten text message, wherein the target text message is the last text message in the historical dialogue information; or, respond to the current user's reply operation based on the candidate feedback text message and perform a corresponding operation based on the reply operation.

[0137] The second receiving and display module 230 is used to receive and display a default recommended text message sent by the server when there is historical chat information in the instant messaging chat interface of the client. The default recommended text message is generated by the server based on the role type and order status information of the client.

[0138] The sending module 240 is configured to send the target recommended text message to the other party's client in response to the target recommended text message selected by the current user in the default recommended text message.

[0139] According to the ride-hailing driver-passenger negotiation assistance device provided in this application embodiment, when there is historical dialogue information in the client's instant messaging dialogue interface, if the client is the sending client of the target text message, the sending client receives and displays the candidate rewritten text message sent by the server. Then, the sending client responds to the selection of the current user (i.e., the sender) and replaces the target text message in the IM dialogue interface with the rewritten text message selected by the current user, thereby realizing the rewriting of the last text message in the sent IM dialogue interface. At the same time, if the client is the receiving client of the target text message, the receiving client receives and displays the candidate feedback text message sent by the server. Then, the receiving client responds to the reply operation of the current user (i.e., the receiver) and performs the corresponding operation, thereby realizing the intelligent reply of the receiving client. This solution, through intelligent rewriting of the target text message and / or intelligent generation of feedback text messages for the target text message, makes the communication measures between the driver and passenger more appropriate and efficient, avoids the problem of negotiation breakdown due to inappropriate wording or poor communication methods, improves communication efficiency and user experience, and increases the platform's order success rate.

[0140] Furthermore, when there is no historical conversation information in the client's instant messaging interface, the client (including the driver's client and the passenger's client) will receive and display the default recommended text message sent by the server. Then, the client (driver's client or passenger's client) responds to the current user's selection and sends the target recommended text message to the other client (corresponding passenger's client or driver's client). This solution can achieve efficient and smooth communication between the driver and passenger during the initial conversation.

[0141] The ride-hailing driver-passenger negotiation assistance device provided in this application embodiment can achieve... Figure 5 The various processes implemented in the method implementation examples will not be described again here to avoid repetition.

[0142] It should be noted that the ride-hailing driver-passenger negotiation assistance device in this application embodiment can be an electronic device or a component of an electronic device, such as an integrated circuit or a chip. The electronic device can be a terminal or other devices besides a terminal. For example, the electronic device can be a mobile phone, tablet computer, laptop computer, PDA, in-vehicle electronic device, mobile internet device (MID), augmented reality (AR) / virtual reality (VR) device, robot, wearable device, ultra-mobile personal computer (UMPC), netbook, or personal digital assistant (PDA), etc. It can also be a server, network attached storage (NAS), personal computer (PC), television (TV), ATM, or self-service machine, etc. This application embodiment does not specifically limit the scope of the device.

[0143] The ride-hailing driver-passenger negotiation assistance device in this application embodiment can be a device with an operating system. This operating system can be a Microsoft (Windows) operating system, an Android operating system, an iOS operating system, or other possible operating systems; this application embodiment does not specifically limit it.

[0144] In some embodiments, such as Figure 8 As shown, this application embodiment also provides an electronic device 800, including a processor 801, a memory 802, and a computer program stored on the memory 802 and executable on the processor 801. When the program is executed by the processor 801, it implements the various processes of the above-described online ride-hailing driver-passenger negotiation assistance method embodiment and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0145] It should be noted that the electronic devices in the embodiments of this application include the mobile electronic devices and non-mobile electronic devices described above.

[0146] This application also provides a non-transitory computer-readable storage medium storing a computer program. When the computer program is executed by a processor, it implements the various processes of the above-described online ride-hailing driver-passenger negotiation assistance method embodiments and achieves the same technical effect. To avoid repetition, it will not be described again here.

[0147] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0148] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the above-described ride-hailing driver-passenger negotiation assistance method.

[0149] The processor is the processor in the electronic device described in the above embodiments. The readable storage medium includes computer-readable storage media, such as computer read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0150] This application embodiment also provides a chip, which includes a processor and a communication interface. The communication interface and the processor are coupled. The processor is used to run programs or instructions to implement the various processes of the above-described online ride-hailing driver-passenger negotiation assistance method embodiment, and can achieve the same technical effect. To avoid repetition, it will not be described again here.

[0151] It should be understood that the chip mentioned in the embodiments of this application may also be referred to as a system-on-a-chip, system chip, chip system, or system-on-a-chip, etc.

[0152] 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.

[0153] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms. Of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the related technology, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk) and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0154] 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.

[0155] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "illustrative embodiment," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0156] Although embodiments of this application have been shown and described, those skilled in the art will understand that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of this application, the scope of which is defined by the claims and their equivalents.

Claims

1. A method for assisting in negotiation between ride-hailing drivers and passengers, characterized in that, Applied to a server, the method includes: Collect basic information about the current order when the current user enters the instant messaging chat interface. The basic information includes at least the order identifier information of the current order. When it is determined that there is historical dialogue information in the instant messaging interface based on the order identification information, the target text message, the sending client, the receiving client and the role type of the target text message in the historical dialogue information are obtained, and the target text message is the last text message in the historical dialogue information. A rewritten text message is generated based on the role type of the sending client, and the rewritten text message is sent to the sending client so that the sending client replaces the target text message with the rewritten text message, and / or a feedback text message of the target text message is generated based on the role type of the receiving client and recommended to the receiving client.

2. The method according to claim 1, characterized in that, The step of generating a rewritten text message based on the role type of the sending client and sending the rewritten text message to the sending client, so that the sending client replaces the target text message with the rewritten text message, includes: The text content of the target text message, the historical dialogue information, and the role type are encapsulated into a rewrite request, and the rewrite request is sent to the large language model. The rewrite request is used to call the large language model to generate candidate rewritten text messages. The system receives the candidate rewritten text messages generated by the large language model and pushes them to the sending client. The sending client then responds to the current user's selection within the candidate rewritten text messages by replacing the target text message with the rewritten text message and clearing other candidate rewritten text messages. The rewritten text message is the text message selected by the current user within the candidate rewritten text messages, and the other candidate rewritten text messages are candidate rewritten text messages other than the original rewritten text message.

3. The method according to claim 2, characterized in that, The step of encapsulating the text content of the target text message, the historical dialogue information, and the role type into a rewrite request includes: Extract the tag information of the target text message, the tag information being used to indicate whether the target text message is text information generated by the large language model; The text content of the target text message, the historical dialogue information, the role type, and the tag information are encapsulated into a rewrite request.

4. The method according to claim 2, characterized in that, While the large language model generates the candidate rewritten text message, the method further includes: A first prompt message is generated and displayed within a first preset range of the target text message. The first prompt message is used to indicate that the target text message is being rewritten. After replacing the target text message with the rewritten text message, the method further includes: A second prompt message is generated and displayed within a second preset range of the rewritten text message. The second prompt message is used to indicate that the target text message has been rewritten. While pushing the candidate rewritten text message to the sending client, the method further includes: Generate a first virtual button for each of the candidate rewritten text messages, and configure the first virtual button to be displayed within a third preset range of the corresponding candidate rewritten text message. The first virtual button is used to respond to the current user's selection of the candidate rewritten text message.

5. The method according to claim 1, characterized in that, The basic information also includes order status information. The step of generating a feedback text message based on the receiving client's role type and recommending it to the receiving client includes: The historical dialogue information, the role type, and the order status information are encapsulated into a prompt request, and the prompt request is sent to the large language model. The prompt request is used to call the large language model to generate candidate feedback text messages. The system receives the candidate feedback text message generated by the large language model and pushes the candidate feedback text message to the receiving client, so that the receiving client can respond to the current user's reply operation based on the candidate feedback text message.

6. The method according to claim 5, characterized in that, While the large language model generates candidate feedback text messages, the method also includes: A third prompt message is generated and configured to be displayed within a fourth preset range of the instant messaging dialog interface. The third prompt message is used to indicate that the candidate feedback text message is being generated. When the candidate feedback text message is a candidate reply text message, receiving the candidate feedback text message generated by the large language model and pushing the candidate feedback text message to the receiving client includes: Generate a second virtual button for each of the candidate feedback text messages, and configure the second virtual button to be displayed within a fifth preset range of the corresponding candidate feedback text message. The second virtual button is used to respond to the current selection of the candidate feedback text message. While pushing the candidate feedback text message to the receiving client, the method further includes: The size of the first floating window is adjusted according to all the candidate feedback text messages so that all the candidate feedback text messages are displayed within the first floating window.

7. The method according to claim 5, characterized in that, The large language model generates the candidate feedback text messages in a streaming manner; The large language model generates the candidate feedback text message in seconds.

8. The method according to claim 1, characterized in that, The basic information also includes order status information, and the method further includes: When it is determined from the order identification information that there is no historical chat information in the instant messaging chat interface, a default recommended text message is obtained based on the current user's role type and the order status information. The default recommended text message is pushed to the current user's client, so that the client responds to the target recommended text message selected by the current user in the default recommended text message and sends the target recommended text message to the other party's client; wherein, the client includes a sending client and a receiving client.

9. The method according to claim 8, characterized in that, While pushing the default recommended text message to the current user's client, the method further includes: A fourth prompt message is generated and configured within a sixth preset range of the second floating window. The second floating window is used to display the default recommended text message, and the fourth prompt message is used to indicate the title information of the second floating window. A third virtual button is generated for each of the default recommended text messages, and the third virtual button is configured to be displayed within a seventh preset range of the corresponding default recommended text message. The third virtual button is used to respond to the current user's selection of the default recommended text message.

10. The method according to any one of claims 1-9, characterized in that, In the process of generating the rewritten text message, or generating the feedback text message, or obtaining the default recommended text message, the method further includes: The system continuously monitors the instant messaging interface, and when a new target text message is detected, it interrupts the process of generating the rewritten text message, generating the feedback text message, or obtaining the default recommended text message, clears the currently displayed generated content, and returns to the steps of obtaining the target text message, the sending client, the receiving client, and the role type of the target text message from the historical dialogue information. The generated content includes at least one of the following: rewritten text messages, feedback text messages, or default recommended text messages during the generation process.

11. The method according to any one of claims 1-9, characterized in that, The method further includes: When the current user exits the instant messaging interface, the current AI recommendation state is not saved. When the current user re-enters the instant messaging chat interface, a target AI type message is obtained. The target AI type message is the last text message of type AI in the historical chat information that was not sent by the current user. Extract the target message content and target message identifier of the target AI type message. Use the target message content and target message identifier to query the target cache. If there is valid AI recommendation content in the target cache, push the AI ​​recommendation content to the current user's client. If there is no valid AI recommendation content in the target cache, encapsulate the target message content and target message identifier into a call request and send it to the large language model. The call request is used to request the large language model to regenerate valid AI recommendation content and push it to the current user's client.

12. The method according to any one of claims 1-9, characterized in that, The method further includes: When an adoption event is detected that the current user has adopted the event, the adoption information of the adoption event is recorded. The adoption event includes at least one of the rewritten text message, the reply text message, and the default recommended text message being adopted by the current user. The adoption information includes at least one of the adoption content, the type of adoption content, the order of adoption content, the order status of adoption content, and the role type of the adopting user. The adopted information is used to iteratively train the large language model.

13. The method according to any one of claims 1-9, characterized in that, The basic information also includes the version information of the current user's client; If the client's version information is a preset version, then the client's instant messaging interface is controlled to display the original text message; if the client's version information is not a preset version, then the client's instant messaging interface is controlled to display the AI ​​text message. The original text message is the text message edited by the current user, and the AI ​​text message is a text message generated based on a large language model, including at least one of rewritten text messages, feedback text messages, and default recommended text messages.

14. A method for assisting in negotiation between ride-hailing drivers and passengers, characterized in that, Applied to a client, the method includes: When historical conversation information exists in the instant messaging interface of the client, the server receives and displays candidate rewritten text messages or candidate feedback text messages, wherein the candidate rewritten text messages and the candidate feedback text messages are generated by the server based on the role type of the client. In response to the current user's selection of a rewritten text message from the candidate rewritten text messages, the target text message is replaced with the rewritten text message, wherein the target text message is the last text message in the historical dialogue information; or In response to the current user's reply operation based on the candidate feedback text message, and based on the reply operation, perform the corresponding operation; When historical chat information exists in the instant messaging interface of the client, a default recommended text message sent by the server is received and displayed, wherein the default recommended text message is generated by the server based on the client's role type and order status information. In response to the target recommended text message selected by the current user in the default recommended text message, the target recommended text message is sent to the other party's client.

15. The method according to claim 14, characterized in that, The candidate feedback text message includes candidate reply text message and recommendation suggestion text message; The response to the current user's reply operation based on the candidate feedback text message, and the execution of corresponding operations based on the reply operation, includes: If the candidate feedback text message is the candidate reply text message, in response to the reply text message selected by the current user in the candidate reply text message, the reply text message is sent to the other party's client; If the candidate feedback text message is the recommendation suggestion text message, then the recommendation suggestion text message will be displayed in the instant messaging dialog interface.

16. The method according to claim 14, characterized in that, The method further includes: Before replacing the target text message with the rewritten text message, a first prompt message is displayed within a first preset range of the target text message. The first prompt message is used to indicate that the target text message is being rewritten. After the target text message is replaced with the rewritten text message, a second prompt message is displayed within a second preset range of the rewritten text message. The second prompt message is used to indicate that the target text message has been rewritten. While receiving and displaying candidate rewritten text messages sent by the server, a first virtual button is displayed within a third preset range of each candidate rewritten text message according to the server's configuration. The first virtual button is used to respond to the current user's selection of the candidate rewritten text message.

17. The method according to claim 16, characterized in that, The method further includes: Before receiving and displaying the candidate feedback text message sent by the server, a third prompt message is displayed within a fourth preset range of the instant messaging dialog interface. The third prompt message is used to indicate that the candidate feedback text message is being generated. While receiving and displaying candidate feedback text messages sent by the server, a second virtual button is displayed within a fifth preset range corresponding to each candidate feedback text message according to the server's configuration. The second virtual button is used to respond to the current user's selection of the candidate feedback text message. The process of receiving and displaying the candidate feedback text message sent by the server includes: displaying the candidate feedback text message in a first floating window according to the server's adjustment.

18. The method according to claim 14, characterized in that, The method further includes: While receiving and displaying the default recommended text message sent by the server, a fourth prompt message is displayed within a sixth preset range of the second floating window according to the server's configuration, and a third virtual button is displayed within a seventh preset range corresponding to each of the default recommended text messages. The fourth prompt message is used to indicate the title information of the second floating window, the second floating window is used to display the default recommended text message, and the third virtual button is used to respond to the current user's selection of the default recommended text message.

19. A ride-hailing driver-passenger negotiation assistance device, characterized in that, The device includes: The data collection module is used to collect basic information about the current order when the current user enters the instant messaging chat interface. The basic information includes at least the order identification information of the current order. The acquisition module is used to acquire, when it is determined from the order identification information that the current order has historical dialogue information, the target text message, the sending client, the receiving client and the role type of the target text message in the historical dialogue information, wherein the target text message is the last text message in the historical dialogue information; The rewrite feedback module is used to generate a rewritten text message based on the role type of the sending client, and send the rewritten text message to the sending client so that the sending client replaces the target text message with the rewritten text message, and / or generate a feedback text message of the target text message based on the role type of the receiving client and recommend it to the receiving client.

20. A ride-hailing driver-passenger negotiation assistance device, characterized in that, The device includes: The first receiving and display module is used to receive and display candidate rewritten text messages or candidate feedback text messages sent by the server when there is historical dialogue information in the instant messaging dialogue interface of the client. The candidate rewritten text messages and the candidate feedback text messages are generated by the server based on the role type of the client. The replacement execution module is configured to respond to the current user's selection of a rewritten text message from the candidate rewritten text messages and replace the target text message with the rewritten text message, wherein the target text message is the last text message in the historical dialogue information; or, respond to the current user's reply operation based on the candidate feedback text message and perform a corresponding operation based on the reply operation. The second receiving and display module is used to receive and display a default recommended text message sent by the server when there is historical chat information in the instant messaging chat interface of the client. The default recommended text message is generated by the server based on the role type and order status information of the client. The selection and sending module is used to send the target recommended text message to the other party's client in response to the target recommended text message selected by the current user in the default recommended text message.

21. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the ride-hailing driver-passenger negotiation assistance method as described in any one of claims 1-18.

22. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the ride-hailing driver-passenger negotiation assistance method as described in any one of claims 1-18.

Images