Method and system for controlling conversation stage task and conversation direction precursor based on large model sales agent

Abstract

The invention discloses a method and system for controlling conversation stage tasks and conversation direction precursor based on a large model sales agent, and belongs to the technical field of computers. The method comprises the following steps: acquiring a customer demand signal sent by a customer; performing first-stage question and answer interaction according to the customer demand signal to obtain a customer basic condition signal; in the second stage, second-stage question and answer interaction is carried out to obtain the client intention; entering a third stage according to the intention of the customer; and in the third stage, third-stage question and answer interaction is carried out, and reservation information is generated according to the customer basic condition signal and the customer intention. Each stage of the whole dialogue has respective objective content, in each stage, no matter what problem is asked by a client, the problem of the client is firstly solved, then main tasks of the stage are asked, and an assistant can not only solve the current problem in a question and answer mode, but also play due sales functions.

Description

Technical Field

[0001] The present invention relates to the field of computer technology, and in particular to a method and system for controlling dialogue stage tasks and dialogue direction precursors based on a large model sales agent. Background Art

[0002] With the rapid development of large language models and intelligent agent technology, the industry has ushered in self-service robot service assistants based on large models and intelligent agents, including intelligent agent assistants such as customer service assistants, intelligent question-and-answer assistants, and self-service query assistants. Its main functions are to configure intelligent agent tools, use intelligent agent large models to identify user semantics, and then call corresponding background tools to give replies or drive back-end business systems. The assistant described in this invention originates from a new sales model innovation-a sales assistant including marketing functions designed using large model intelligent agents. This assistant needs to deeply understand consumer needs through dialogue, and then recommend small models based on demand understanding. Then the assistant must carry out continuous marketing persuasion work around the recommended items and customer needs, and finally persuade customers to reach an ordering intention. Finally, ask the customer for door-to-door contact information to form a sales closed loop of real customer manager door-to-door service. It aims to provide enterprises with a more efficient, more accurate, and more cost-effective sales and marketing solution.

[0003] In order to realize the above-mentioned sales intelligent service assistant with multiple links and multiple roles, a large language model, intelligent agent, and a small recommendation model at the back end are required, as well as a robot that is not intelligent enough in sales skills and sales guidance and has human-like sales emotional intelligence.

[0004] For example, there is a common statement of rights in the industry:

[0005] An autonomous marketing system based on a large language model multi-agent framework (application number: CN202311561764.3 Publication (announcement) number: CN117495505A). The main points are: first, a review of the traditional agent development method; second, the switching of sales roles is achieved through a multi-agent method. That is, an autonomous marketing system based on a large language model multi-agent framework, including an agent interactive system module, a decentralized autonomous organization module, and an agent network structure module.

[0006] The patent "CN117495505A" advocates a "sales service intelligent agent assistant" that focuses on creating assistants with different roles at different stages based on multiple agents to drive dialogue, but it does not solve the following two key problems. Especially for this marketing and sales scenario, customers will face many, complex and disordered questions. When traditional dialogue robots face many, complex and disordered questions, it will lead to mismatch of intelligent agent tools and lack of role goals, which will cause confusion in context flow and dialogue effect. Even with large-model multi-round dialogue capabilities, it will be of no help.

[0007] In the wave of digitalization and intelligence, the sales industry is facing unprecedented changes. With the rapid development of big models and intelligent agent technology, various application developers have devoted themselves to the research and development of sales assistants, aiming to improve sales efficiency through automation and intelligence. However, most of the existing sales assistants on the market remain at the level of mechanical single-round dialogue, lacking the ability to deeply understand and conduct multi-round dialogues, and cannot truly simulate the sales process of natural persons, thus limiting their application effect in actual sales scenarios. Summary of the invention

[0008] The purpose of the present invention is to provide a method and system for controlling dialogue stage tasks and dialogue direction precursors based on a large-model sales agent, and to build a highly intelligent and humanized sales assistant according to the characteristics of sales dialogues being divided into stages and roles, and simulating the process of natural persons driving dialogues forward and facilitating sales results. By customizing the assistant's global direction variables and the method of controlling and modifying stage background prompts, this sales robot assistant can display different dialogue strategies and direction controls at different sales stages, achieve natural interaction with potential customers, significantly improve sales efficiency, reduce sales costs, and provide customers with a better quality, more personalized sales experience. At the same time, it will also provide a new idea and method for application developers to promote the innovation and development of sales assistant technology.

[0009] In order to solve the above technical problems, the present invention provides a method for controlling dialogue stage tasks and dialogue direction precursors based on a large model sales agent, comprising the following steps:

[0010] Obtain customer demand signals from customers;

[0011] Conduct the first phase of question-and-answer interaction based on the customer demand signal to obtain the customer's basic situation signal; enter the second or third phase based on the customer demand signal and the customer's basic situation signal;

[0012] In the second stage, the second stage question-and-answer interaction is carried out to obtain the customer's intention; according to the customer's intention, the third stage is entered;

[0013] In the third stage, the third stage question-and-answer interaction is carried out to generate appointment information based on the customer's basic situation signals and customer intentions.

[0014] Preferably, the first phase of question-answering interaction is performed according to the customer demand signal to obtain the customer basic situation signal, which specifically includes the following steps:

[0015] Generate introduction signals based on customer demand signals;

[0016] Based on customer demand signals and introduction signals, conduct the first phase of question-and-answer interaction with the customer to obtain basic customer situation signals from the customer.

[0017] Preferably, after obtaining the basic information signal of the customer, the following steps are also included:

[0018] Generate recommendation results based on customer demand signals and customer basic situation signals.

[0019] Preferably, generating a recommendation result according to the customer demand signal and the customer basic situation signal specifically includes the following steps:

[0020] Input customer demand signals and basic information signals into the large model agent to obtain recommendation results.

[0021] Preferably, the second stage of question-and-answer interaction is performed to obtain the customer's intention, which specifically includes the following steps:

[0022] Generate sales pitches based on customer demand signals, customer basic situation signals and recommendation results;

[0023] Based on the sales script, conduct the second phase of question-and-answer interaction with the customer to obtain the customer's intention.

[0024] Preferably, the third stage of question-answer interaction is performed to generate appointment information according to the customer's basic situation signal and the customer's intention, which specifically includes the following steps:

[0025] Conduct the third phase of Q&A interaction with customers based on their intentions;

[0026] Generate inquiry signals based on customer intentions and send them to customers;

[0027] The customer sends a contact signal based on the inquiry signal;

[0028] Generate appointment information based on customer basic information signals, customer intentions and contact information signals.

[0029] Preferably, the customer basic situation signal includes industry, area, number of people and turnover.

[0030] Preferably, the contact information signal includes a contact number, address, identity information and service time.

[0031] The present invention also provides a system for controlling dialogue stage tasks and dialogue direction precursors based on a large model sales agent, comprising:

[0032] An acquisition module, used to acquire customer demand signals sent by customers;

[0033] The first-stage processing module is used to perform the first-stage question-answer interaction according to the customer demand signal to obtain the customer basic situation signal; and enter the second stage or the third stage according to the customer demand signal and the customer basic situation signal;

[0034] The second stage processing module is used to conduct the second stage question-answer interaction in the second stage to obtain the customer's intention; according to the customer's intention, enter the third stage;

[0035] The third stage processing module is used to conduct the third stage question and answer interaction in the third stage and generate appointment information according to the customer's basic situation signals and customer intentions.

[0036] Compared with the prior art, the present invention has the following beneficial effects:

[0037] 1. Each stage of the whole conversation has its own main content. Within each stage, no matter what questions the customer asks, the customer’s problem should be solved first, and then the main task of this stage should be followed up. This allows an assistant to not only solve current problems by answering questions, but also perform due sales functions.

[0038] 2. The control of the direction of the conversation process ensures that the customer will not intentionally or unintentionally trigger the "transition phase" of the previous phase in the subsequent conversation, causing the entire phase to flow back, thereby causing the overall conversation direction to flow backward. The essence is to let the robot simulate the sales forerunner thinking of people in the conversation, which is an effective way to avoid mechanical questions and answers. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The specific implementation modes of the present invention are further described in detail below with reference to the accompanying drawings.

[0040] Figure 1 It is a flow chart of a method for controlling dialogue stage tasks and dialogue direction precursors based on a large model sales agent of the present invention;

[0041] Figure 2 This is a flowchart of a sales assistant based on a large model agent;

[0042] Figure 3 It is a flowchart of the tasks that control the dialogue stage. DETAILED DESCRIPTION

[0043] Many specific details are set forth in the following description to facilitate a full understanding of the present invention. However, the present invention can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the connotation of the present invention, so the present invention is not limited to the specific implementation disclosed below.

[0044] The terms used in one or more embodiments of this specification are only for the purpose of describing specific embodiments, and are not intended to limit one or more embodiments of this specification. The singular forms of "a", "said" and "the" used in one or more embodiments of this specification and the appended claims are also intended to include plural forms, unless the context clearly indicates other meanings. It should also be understood that the term "and / or" used in one or more embodiments of this specification refers to and includes any or all possible combinations of one or more associated listed items.

[0045] It should be understood that although the terms first, second, etc. may be used to describe various information in one or more embodiments of this specification, this information should not be limited to these terms. These terms are only used to distinguish the same type of information from each other. For example, without departing from the scope of one or more embodiments of this specification, the first may also be referred to as the second, and similarly, the second may also be referred to as the first. Depending on the context, the word "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0046] The present invention is further described in detail below in conjunction with the accompanying drawings:

[0047] The present invention provides a method for controlling dialogue stage tasks and dialogue direction precursors based on a large model sales agent, comprising the following steps:

[0048] Obtain customer demand signals from customers;

[0049] Conduct the first phase of question-and-answer interaction based on the customer demand signal to obtain the customer's basic situation signal; enter the second or third phase based on the customer demand signal and the customer's basic situation signal;

[0050] In the second stage, the second stage question-and-answer interaction is carried out to obtain the customer's intention; according to the customer's intention, the third stage is entered;

[0051] In the third stage, the third stage question-and-answer interaction is carried out to generate appointment information based on the customer's basic situation signals and customer intentions.

[0052] Preferably, the first phase of question-answering interaction is performed according to the customer demand signal to obtain the customer basic situation signal, which specifically includes the following steps:

[0053] Generate introduction signals based on customer demand signals;

[0054] Based on customer demand signals and introduction signals, conduct the first phase of question-and-answer interaction with the customer to obtain basic customer situation signals from the customer.

[0055] Preferably, after obtaining the basic information signal of the customer, the following steps are also included:

[0056] Generate recommendation results based on customer demand signals and customer basic situation signals.

[0057] Preferably, generating a recommendation result according to the customer demand signal and the customer basic situation signal specifically includes the following steps:

[0058] Input customer demand signals and basic information signals into the large model agent to obtain recommendation results.

[0059] Preferably, the second stage of question-and-answer interaction is performed to obtain the customer's intention, which specifically includes the following steps:

[0060] Generate sales pitches based on customer demand signals, customer basic situation signals and recommendation results;

[0061] Based on the sales script, conduct the second phase of question-and-answer interaction with the customer to obtain the customer's intention.

[0062] Preferably, the third stage of question-answer interaction is performed to generate appointment information according to the customer's basic situation signal and the customer's intention, which specifically includes the following steps:

[0063] Conduct the third phase of Q&A interaction with customers based on their intentions;

[0064] Generate inquiry signals based on customer intentions and send them to customers;

[0065] The customer sends a contact signal based on the inquiry signal;

[0066] Generate appointment information based on customer basic information signals, customer intentions and contact information signals.

[0067] Preferably, the customer basic situation signal includes industry, area, number of people and turnover.

[0068] Preferably, the contact information signal includes a contact number, address, identity information and service time.

[0069] The present invention also provides a system for controlling dialogue stage tasks and dialogue direction precursors based on a large model sales agent, comprising:

[0070] An acquisition module, used to acquire customer demand signals sent by customers;

[0071] The first-stage processing module is used to perform the first-stage question-answer interaction according to the customer demand signal to obtain the customer basic situation signal; and enter the second stage or the third stage according to the customer demand signal and the customer basic situation signal;

[0072] The second stage processing module is used to conduct the second stage question-answer interaction in the second stage to obtain the customer's intention; according to the customer's intention, enter the third stage;

[0073] The third stage processing module is used to conduct the third stage question and answer interaction in the third stage and generate appointment information according to the customer's basic situation signals and customer intentions.

[0074] In order to better illustrate the technical effect of the present invention, the present invention provides the following specific embodiments to illustrate the above technical process:

[0075] Embodiment 1: A method for controlling dialogue stage tasks and dialogue direction precursors based on a large model sales agent, comprising:

[0076] 1. Sales assistant configuration scheme and principle based on large model agent, such as Figure 2 As shown;

[0077] Generally, the configuration and development process based on the large model agent is to first configure the assistant, and then register the interface developed by the business to complete the configuration of the tool. In the running state, the solution process is as follows:

[0078] 1. Assistant Configuration Configure the sales service assistant's description, opening remarks, and parameters such as temperature values ​​related to the large model.

[0079] 2. Tool development and tool configuration

[0080] If the assistant needs to call business system interfaces such as product recommendations, appointment ordering, etc. during the conversation, it is necessary to first develop such interfaces using traditional IT means, then register them to the tool management module of the chat front-line service platform, and finally incorporate them into the assistant's tool configuration.

[0081] 3. In addition to relying on the intelligent agent matching tool to communicate during the conversation process, ordinary conversation tasks in the process need to rely on the meaning understanding and text generation of the large model itself, as well as the ability to conduct multiple rounds of conversations. The disadvantage is that many questions are prone to over-generalization, and it is impossible to actively guide customers to achieve their goals at the current stage during the answer.

[0082] Second, the tasks of controlling the dialogue stage, such as Figure 3 As shown;

[0083] An excellent and natural sales service assistant should have the awareness of natural person sales, that is, in order to achieve sales goals, the communication process with customers needs to be naturally divided into stages. Each stage has its own dialogue task goals and has considerable autonomy and purposefulness.

[0084] The sales assistant is divided into three stages: demand contact, recommendation promotion, and contact information acquisition.

[0085] 1. Configuration of key tools for transition phase.

[0086] The key question and answer for transitioning from the first stage to the second stage is the conversation that triggers the product recommendation tool. The transition tool from the second to the third stage is the intention achievement tool.

[0087] 2. Tool background prompt word configuration.

[0088] In each stage, there are background prompts, that is, no matter what questions the customer asks in this stage, the assistant will first solve the question, and then ask the current stage of the dialogue task through the big model. This function is mainly completed through background prompts.

[0089] 3. All questions in the process are summarized by adding the background prompt word to the big model, so that the big model has the ability to automatically ask or automatically complete the answer to the current question. And only one background prompt word runs at the same time in a session.

[0090] 3. Control the direction of the conversation

[0091] During the conversation, there may be situations where the customer's question jumps to another stage. For example, if the customer has reached an intention to order a product, the assistant will ask for the customer's contact information in the third stage, but the customer will answer about his or her basic situation. At this time, the assistant will hit the "product recommendation tool" of the transition stage according to the tool matching principle, and start a conversation based on the background prompt words of the second stage. In order to avoid this kind of stage backtracking or jumping without direction control, a solution for forward control of the assistant conversation is designed to make the assistant directional.

[0092] 1. Assign a stage ID to each dialogue stage. For example, this assistant will mark demand contact, recommendation promotion, and contact information acquisition as 1, 2, and 3 respectively.

[0093] 2. Control the forward control of the assistant by setting a global unique variable in the session. Set this global variable when each assistant session is opened. The initial default is the first stage. When it reaches different stages, it is assigned the representative variable of the stage. However, when a new stage is triggered, when the global variable is about to be modified, the size of the global variable and the target stage identifier must be judged. As long as the newly triggered stage identifier is less than or equal to the current global variable, the global variable is not allowed to be updated; otherwise, the global variable is allowed to be updated. In this way, the entire sales conversation can only move forward, solving the problem that the sales assistant jumps to different stages or even backtracks due to multiple factors of questions and answers, resulting in confusion in the conversation process.

[0094] Example 1:

[0095] During the pre-sales phase for government and enterprise business opportunities, a certain operator opened up the "Sales Service Assistant" through the government and enterprise self-service platform for small and medium-sized enterprise customers to conduct initial customer contact and explore business opportunities, which achieved good business opportunity marketing results. At the same time, it achieved the goal of reducing the workload of account managers and training new account managers.

[0096] In the several embodiments provided by the present invention, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic, for example, the division of the modules, modules or units is only a logical function division, and there may be other division methods in actual implementation, for example, multiple units, modules or components can be combined or integrated into another device, or some features can be ignored or not executed.

[0097] The units may or may not be physically separated, and the components displayed as units may be one physical unit or multiple physical units, that is, they may be located in one place or distributed in multiple different places. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0098] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0099] In particular, according to the embodiments disclosed in the present invention, the process described above with reference to the flowchart can be implemented as a computer software program. For example, the embodiments disclosed in the present invention include a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from the network through the communication part, and / or installed from a removable medium. When the computer program is executed by the central processing unit (CPU), the above-mentioned functions defined in the method of the present invention are executed. It should be noted that the above-mentioned computer-readable medium of the present invention can be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium can be, for example, but not limited to, a system, device or device of an electrical, magnetic, optical, electromagnetic, infrared segment, or semiconductor, or any combination of the above.

[0100] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each square box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the module, the program segment or a part of the code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the square box can also occur in a sequence different from that marked in the accompanying drawings. For example, two square boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each square box in the block diagram and / or flow chart, and the combination of the square boxes in the block diagram and / or flow chart can be implemented with a dedicated hardware-based system that performs the specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0101] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention should be based on the protection scope of the claims.

Claims

1. A method for controlling dialogue stage tasks and dialogue direction precursors based on a large model sales agent, characterized in that: The following steps are involved: Obtain customer demand signals from customers; Conduct the first phase of question-and-answer interaction based on the customer demand signal to obtain the customer's basic situation signal; enter the second or third phase based on the customer demand signal and the customer's basic situation signal; In the second stage, the second stage question-and-answer interaction is carried out to obtain the customer's intention; according to the customer's intention, the third stage is entered; In the third stage, the third stage question-and-answer interaction is carried out to generate appointment information based on the customer's basic situation signals and customer intentions.

2. The method for controlling dialogue stage tasks and dialogue direction precursors based on a large model sales agent according to claim 1, characterized in that: The first phase of question-and-answer interaction is performed based on the customer demand signal to obtain the customer's basic situation signal, which specifically includes the following steps: Generate introduction signals based on customer demand signals; Based on customer demand signals and introduction signals, conduct the first phase of question-and-answer interaction with the customer to obtain basic customer situation signals from the customer.

3. The method for controlling the dialogue stage tasks and the dialogue direction precursor based on the large model sales agent according to claim 2 is characterized in that: After obtaining the basic information of the customer, the following steps are also included: Generate recommendation results based on customer demand signals and customer basic situation signals.

4. The method for controlling dialogue stage tasks and dialogue direction precursors based on a large model sales agent according to claim 3, characterized in that: Generate recommendation results based on customer demand signals and customer basic situation signals, which specifically includes the following steps: Input customer demand signals and basic information signals into the large model agent to obtain recommendation results.

5. The method for controlling dialogue stage tasks and dialogue direction precursors based on a large model sales agent according to claim 4, characterized in that: Conduct the second phase of question-and-answer interaction to obtain customer intentions, which includes the following steps: Generate sales pitches based on customer demand signals, customer basic situation signals and recommendation results; Based on the sales script, conduct the second phase of question-and-answer interaction with the customer to obtain the customer's intention.

6. The method for controlling dialogue stage tasks and dialogue direction precursors based on a large model sales agent according to claim 5, characterized in that: The third stage of question-and-answer interaction is carried out to generate appointment information based on the customer's basic situation signals and customer intentions, which specifically includes the following steps: Conduct the third phase of Q&A interaction with customers based on their intentions; Generate inquiry signals based on customer intentions and send them to customers; The customer sends a contact signal based on the inquiry signal; Generate appointment information based on customer basic information signals, customer intentions and contact information signals.

7. The method for controlling dialogue stage tasks and dialogue direction precursors based on a large model sales agent according to claim 6, characterized in that: The customer basic situation signals include industry, area, number of people and turnover.

8. The method for controlling dialogue stage tasks and dialogue direction precursors based on a large model sales agent according to claim 7, characterized in that: The contact information signal includes contact number, address, identity information and service time.

9. A system for controlling dialogue stage tasks and dialogue direction precursors based on a large model sales agent, used to implement the method for controlling dialogue stage tasks and dialogue direction precursors based on a large model sales agent as described in any one of claims 1 to 8, characterized in that: include: An acquisition module, used to acquire customer demand signals sent by customers; The first-stage processing module is used to perform the first-stage question-answer interaction according to the customer demand signal to obtain the customer basic situation signal; and enter the second stage or the third stage according to the customer demand signal and the customer basic situation signal; The second stage processing module is used to conduct the second stage question-answer interaction in the second stage to obtain the customer's intention; according to the customer's intention, enter the third stage; The third stage processing module is used to conduct the third stage question and answer interaction in the third stage and generate appointment information according to the customer's basic situation signals and customer intentions.

Images