A data processing method, system and device for assisting in expanding customers

Abstract

The embodiment of the invention discloses a data processing method, system and device for assisting customer extension. The method comprises the steps of obtaining query information of a user, obtaining priori knowledge corresponding to the query information, generating an Internet extension sub-task, a knowledge base extension sub-task and a label extension sub-task according to the priori knowledge and the query information, and respectively calling extension agents corresponding to the sub-tasks to generate an intermediate extension result. And generating an auxiliary customer extension result according to the intermediate customer extension result through the summarizing agent and outputting the auxiliary customer extension result. Therefore, multiple customer extension means can be integrated through artificial intelligence to provide intelligent assistance for customer extension, data fusion and collaborative analysis are realized, the cost is reduced, the customer extension depth and breadth are improved, and the conversion rate is further improved.

Description

(19) State Intellectual Property Office (12) Invention Patent Application (10) Application Publication Number (43) Application Publication Date (21) Application Number 202511554594.5 (22) Application Date 2025.10.28 (71) Applicant Alibaba (China) Co., Ltd. Address 310052, Room 508, 5th Floor, Building 4, No. 699, Wangshang Road, Changhe Street, Binjiang District, Hangzhou, Zhejiang Province (72) Inventors Li Xuanxu, He Wendong, Bai Hao, Gao Yirui, Chen Ronglong, Gu Shuo, Ding Yifei (74) Patent Agency Beijing Ruipai Intellectual Property Agency Co., Ltd. 11597 Patent Attorney Liu Feng, Wang Baofu (51) Int.Cl. G06Q 30 / 0201 (2023.01) G06N 3 / 006 (2023.01) (54) Invention Title A Data Processing Method, System and Device for Assisting Customer Acquisition (57) Abstract The embodiments of the present invention disclose a data processing method, system and device for assisting customer acquisition. By acquiring user query information and corresponding prior knowledge, and generating internet-based customer acquisition sub-tasks, knowledge base-based customer acquisition sub-tasks, and tag-based customer acquisition sub-tasks based on the prior knowledge and query information, each sub-task's corresponding customer acquisition agent generates intermediate customer acquisition results. Finally, a summary agent generates and outputs auxiliary customer acquisition results based on these intermediate results. Thus, artificial intelligence can be used to integrate multiple customer acquisition methods to provide intelligent assistance, achieving data fusion and collaborative analysis, reducing costs, increasing the depth and breadth of customer acquisition, and ultimately improving conversion rates. Claims 2 pages, Description 13 pages, Drawings 8 pages, CN 121526660 A 2026.02.13 CN 1 21 52 66 60 A 1. A data processing method for assisting customer acquisition, characterized in that the method includes: obtaining user query information; obtaining prior knowledge corresponding to the query information; generating sub-tasks based on the prior knowledge and the query information, the sub-tasks including internet customer acquisition sub-tasks, knowledge base customer acquisition sub-tasks, and tag customer acquisition sub-tasks; respectively calling the customer acquisition agents corresponding to each sub-task to generate intermediate customer acquisition results; and generating and outputting auxiliary customer acquisition results based on the intermediate customer acquisition results through a summarizing agent. 2. The method according to claim 1, characterized in that obtaining the prior knowledge corresponding to the query information includes: calling an internet retrieval link to obtain background knowledge corresponding to the query information; generating the prior knowledge based on the background knowledge, the prior knowledge including one or more of concept definitions, problem objectives, and analysis frameworks. 3. The method according to claim 1, wherein the customer acquisition agent generates intermediate customer acquisition results by one or more of the following: keyword expansion, query information analysis, background knowledge and analysis suggestions, and output result analysis.4. The method according to claim 1, wherein the customer acquisition intelligence agent includes an internet customer acquisition intelligence agent, a knowledge base customer acquisition intelligence agent, and a tag customer acquisition intelligence agent; wherein, the step of calling the customer acquisition intelligence agents corresponding to each sub-task to generate intermediate customer acquisition results includes: executing the internet customer acquisition sub-task through the internet customer acquisition intelligence agent to generate internet customer acquisition results, wherein the internet customer acquisition results are Markdown customer acquisition reports; executing the knowledge base customer acquisition sub-task through the knowledge base customer acquisition intelligence agent to generate knowledge base customer acquisition results, wherein the knowledge base customer acquisition results are search conditions; executing the tag customer acquisition sub-task through the tag customer acquisition intelligence agent to generate tag customer acquisition results, wherein the tag customer acquisition results are DSL search statements. 5. The method according to claim 4, wherein the step of generating and outputting auxiliary customer acquisition results based on the intermediate customer acquisition results through the summarizing intelligence agent includes: using a predetermined recognition model to recognize the Markdown customer acquisition report to extract enterprise information; using the search conditions to obtain the hit paragraphs and obtaining the enterprise information corresponding to the hit paragraphs; using the DSL search statement to retrieve the corresponding enterprise information; and summarizing the obtained enterprise information to obtain the auxiliary customer acquisition results. 6. The method according to any one of claims 1-5, characterized in that the method further comprises: outputting the events of the customer acquisition agent and the summarizing agent. 7. A data processing system for assisting customer acquisition, characterized in that the system comprises: a planner, configured to, after obtaining user query information, obtain prior knowledge corresponding to the query information, and generate sub-tasks based on the prior knowledge and the query information, the sub-tasks including internet customer acquisition sub-tasks, knowledge base customer acquisition sub-tasks, and tag customer acquisition sub-tasks; multiple customer acquisition agents, configured to execute corresponding sub-tasks to obtain intermediate customer acquisition results; and a summarizing agent, configured to generate and output auxiliary customer acquisition results based on the intermediate customer acquisition results. 8. The system according to claim 7, characterized in that the planner is configured to invoke an internet retrieval link to obtain background knowledge corresponding to the query information, and generate the prior knowledge based on the background knowledge, the prior knowledge including one or more of concept definitions, problem objectives, and analysis frameworks. 9. The system according to claim 7, wherein the customer acquisition agent generates intermediate customer acquisition results by one or more of keyword expansion, query information analysis, background knowledge and analysis suggestions, and output result analysis. 10. The system according to claim 9, wherein the customer acquisition agent comprises an internet customer acquisition agent, a knowledge base customer acquisition agent, and a tag customer acquisition agent; the internet customer acquisition agent executes the internet customer acquisition sub-task to generate internet customer acquisition results, and the internet customer acquisition results are Markdown customer acquisition reports; the knowledge base customer acquisition agent…The system executes the knowledge base customer acquisition subtask to generate knowledge base customer acquisition results, which serve as search criteria; the tag customer acquisition agent executes the tag customer acquisition subtask to generate tag customer acquisition results, which serve as DSL search statements. 11. The system according to claim 10, wherein the summarizing agent is used to identify the Markdown customer acquisition report using a predetermined recognition model to extract enterprise information; uses the search criteria to obtain the matched paragraphs, and obtains the enterprise information corresponding to the matched paragraphs; uses the DSL search statement to retrieve the corresponding enterprise information; and summarizes the obtained enterprise information to obtain the auxiliary customer acquisition results. 12. A data processing device for assisting customer acquisition, characterized in that the device comprises: a query information acquisition unit for acquiring user query information; a prior knowledge acquisition unit for acquiring prior knowledge corresponding to the query information; a subtask generation unit for generating subtasks based on the prior knowledge and the query information, the subtasks including internet customer acquisition subtasks, knowledge base customer acquisition subtasks, and tag customer acquisition subtasks; an intermediate customer acquisition result generation unit for respectively calling the customer acquisition agent corresponding to each subtask to generate intermediate customer acquisition results; and an auxiliary customer acquisition result generation unit for generating and outputting auxiliary customer acquisition results based on the intermediate customer acquisition results through a summarizing agent. 13. An electronic device, comprising a memory and a processor, characterized in that the memory is used to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor to implement the method as described in any one of claims 1-6. 14. A computer-readable storage medium storing computer program instructions thereon, characterized in that the computer program instructions, when executed by a processor, implement the method as described in any one of claims 1-6. Claims 2 / 2 Page 3 CN 121526660 A A Data Processing Method, System and Device for Assisting Customer Acquisition Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a data processing method, system and device for assisting customer acquisition. Background Art

[0002] With the intensification of market competition, enterprises are increasingly demanding higher accuracy and real-time performance in customer outreach. Efficient and intelligent customer acquisition mechanisms have become an important part of the core competitiveness of enterprises. Existing technologies mainly rely on manual screening methods for customer acquisition, such as telephone marketing and ground promotion, which have problems such as high labor costs and low conversion rates. At the same time, enterprises generally use static tag systems to classify customers, and internal knowledge bases and external Internet data form data silos, making it difficult to achieve data fusion and collaborative analysis, thus limiting the depth and breadth of customer insights. Summary of the Invention

[0003] In view of this, the purpose of the embodiments of the present invention is to provide a data processing method, system and device for assisting customer acquisition.The device can integrate multiple customer acquisition methods through artificial intelligence to provide intelligent assistance for customer acquisition, realize data fusion and collaborative analysis, reduce costs, improve the depth and breadth of customer acquisition, and thus improve conversion rate.

[0004] In a first aspect, embodiments of the present invention provide a customer acquisition method, the method comprising: obtaining user query information; obtaining prior knowledge corresponding to the query information; generating sub-tasks based on the prior knowledge and the query information, the sub-tasks including internet customer acquisition sub-tasks, knowledge base customer acquisition sub-tasks and tag customer acquisition sub-tasks; respectively calling the customer acquisition intelligent agent corresponding to each sub-task to generate intermediate customer acquisition results; and generating auxiliary customer acquisition results based on the intermediate customer acquisition results through a summarizing intelligent agent and outputting them.

[0005] In some embodiments, obtaining the prior knowledge corresponding to the query information includes: calling the internet retrieval link to obtain background knowledge corresponding to the query information; generating the prior knowledge based on the background knowledge, the prior knowledge including one or more of concept definitions, problem objectives and analysis frameworks.

[0006] In some embodiments, the customer acquisition agent generates intermediate customer acquisition results by one or more of the following: keyword expansion, query information analysis, background knowledge and analysis suggestions, and output result analysis.

[0007] In some embodiments, the customer acquisition agent includes an internet customer acquisition agent, a knowledge base customer acquisition agent, and a tag customer acquisition agent; wherein, the step of calling the customer acquisition agents corresponding to each subtask to generate intermediate customer acquisition results includes: executing the internet customer acquisition subtask through the internet customer acquisition agent to generate internet customer acquisition results, wherein the internet customer acquisition results are Markdown customer acquisition reports; executing the knowledge base customer acquisition subtask through the knowledge base customer acquisition agent to generate knowledge base customer acquisition results, wherein the knowledge base customer acquisition results are search conditions; executing the tag customer acquisition subtask through the tag customer acquisition agent to generate tag customer acquisition results, wherein the tag customer acquisition results are DSL search statements.

[0008] In some embodiments, the step of generating and outputting auxiliary customer acquisition results by the summarizing agent based on the intermediate customer acquisition results includes: using a predetermined recognition model to recognize the Markdown customer acquisition report to extract enterprise information; using the search conditions to obtain the hit paragraphs and obtain the enterprise information corresponding to the hit paragraphs; using the DSL search statement to retrieve the corresponding enterprise information; and summarizing the obtained enterprise information to obtain the auxiliary customer acquisition results.

[0009] In some embodiments, the method further includes: outputting the events of the customer acquisition agent and the summarizing agent.

[0010] In a second aspect, embodiments of the present invention provide a customer acquisition system, the system including: a planner, used to obtain prior knowledge corresponding to the query information after obtaining the user's query information, andSubtasks are generated based on the prior knowledge and the query information. These subtasks include internet customer acquisition subtasks, knowledge base customer acquisition subtasks, and tag-based customer acquisition subtasks. Multiple customer acquisition agents are used to execute corresponding subtasks to obtain intermediate customer acquisition results. A summarizing agent is used to generate auxiliary customer acquisition results based on the intermediate customer acquisition results and output them.

[0011] In some embodiments, the planner is used to invoke internet retrieval links to obtain background knowledge corresponding to the query information, and generate the prior knowledge based on the background knowledge. The prior knowledge includes one or more of concept definitions, problem objectives, and analysis frameworks.

[0012] In some embodiments, the customer acquisition agents generate intermediate customer acquisition results by including one or more of keyword expansion, query information analysis, background knowledge and analysis suggestions, and output result analysis.

[0013] In some embodiments, the customer acquisition agent includes an internet customer acquisition agent, a knowledge base customer acquisition agent, and a tag customer acquisition agent. The internet customer acquisition agent executes the internet customer acquisition sub-task to generate internet customer acquisition results, which are Markdown customer acquisition reports. The knowledge base customer acquisition agent executes the knowledge base customer acquisition sub-task to generate knowledge base customer acquisition results, which are search conditions. The tag customer acquisition agent executes the tag customer acquisition sub-task to generate tag customer acquisition results, which are DSL search statements.

[0014] In some embodiments, the summarizing agent is used to identify the Markdown customer acquisition report using a predetermined recognition model to extract enterprise information; use the search conditions to obtain the hit paragraphs and obtain the enterprise information corresponding to the hit paragraphs; use the DSL search statement to retrieve the corresponding enterprise information; and summarize the obtained enterprise information to obtain the auxiliary customer acquisition results.

[0015] In a third aspect, embodiments of the present invention provide a customer acquisition device, the device comprising: a query information acquisition unit, configured to acquire query information from a user; a prior knowledge acquisition unit, configured to acquire prior knowledge corresponding to the query information; a subtask generation unit, configured to generate subtasks based on the prior knowledge and the query information, the subtasks including internet customer acquisition subtasks, knowledge base customer acquisition subtasks, and tag customer acquisition subtasks; an intermediate customer acquisition result generation unit, configured to call the customer acquisition agent corresponding to each subtask to generate intermediate customer acquisition results; and an auxiliary customer acquisition result generation unit, configured to generate auxiliary customer acquisition results based on the intermediate customer acquisition results through a summarizing agent and output them.

[0016] In a fourth aspect, embodiments of the present invention provide an electronic device, comprising a memory and a processor, the memory being configured to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor to implement the method described in the first aspect.

[0017] In a fifth aspect, embodiments of the present invention provide a computer-readable storage medium storing computer program instructions thereon, which, when executed by a processor, implement the method described in the first aspect.

[0018] The technical solution of the embodiments of the present invention obtains user query information, obtains prior knowledge corresponding to the query information, generates Internet customer acquisition sub-tasks, knowledge base customer acquisition sub-tasks, and tag customer acquisition sub-tasks based on the prior knowledge and query information, respectively calls the customer acquisition intelligent agent corresponding to each sub-task to generate intermediate customer acquisition results, and summarizes the intelligent agent to generate auxiliary customer acquisition results based on the intermediate customer acquisition results and outputs them. Thus, multiple customer acquisition methods can be integrated through artificial intelligence to provide intelligent assistance for customer acquisition, realize data fusion and collaborative analysis, reduce costs, improve the depth and breadth of customer acquisition, and thereby improve conversion rates.

[0019] The above and other objects, features and advantages of the present invention will become clearer from the following description of embodiments of the present invention with reference to the accompanying drawings, in which: FIG1 is a schematic diagram of a customer acquisition system according to an embodiment of the present invention; FIG2 is a flowchart of a customer acquisition method according to an embodiment of the present invention; FIG3 is a flowchart of acquiring prior knowledge according to an embodiment of the present invention; FIG4 is a flowchart of acquiring intermediate customer acquisition results according to an embodiment of the present invention; FIG5 is a flowchart of generating auxiliary customer acquisition results according to an embodiment of the present invention; FIG6 is a schematic diagram of the customer acquisition task execution process according to an embodiment of the present invention; FIG7 is a schematic diagram of a human-computer interaction interface according to an embodiment of the present invention; FIG8 is a schematic diagram of a customer acquisition device according to an embodiment of the present invention; FIG9 is a schematic diagram of an electronic device according to an embodiment of the present invention. Detailed Description

[0020] The present application is described below based on embodiments, but the present application is not limited to these embodiments. In the following detailed description of the present application, some specific details are described in detail. The present application can be fully understood by those skilled in the art without these details. In order to avoid obscuring the substance of the present application, well-known methods, processes, flows, elements and circuits are not described in detail.

[0021] Furthermore, those skilled in the art should understand that the accompanying drawings provided herein are for illustrative purposes and are not necessarily drawn to scale.

[0022] Unless the context explicitly requires it, the terms "comprising," "including," and similar words throughout the application should be interpreted as including rather than exclusive or exhaustive; that is, meaning "including but not limited to."

[0023] In the description of this application, it should be understood that the terms "first," "second," etc., are used for descriptive purposes only and should not be construed as indicating or implying relative importance. Furthermore, in the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0024] The solutions described in this specification and embodiments, if involving personal information processing, will all be based on legality.Processing will be carried out only within the scope stipulated or agreed upon, provided that the personal information subject consents (e.g., with the consent of the personal information subject, or as necessary for the performance of a contract). Users' refusal to process personal information beyond the necessary information for basic functions will not affect their use of basic functions.

[0025] In today's highly digitalized and competitive business environment, customer acquisition capabilities have become a core driver of sustainable corporate growth. As the market shifts from extensive expansion to refined operation, companies have higher requirements for the depth of customer insights, the accuracy of outreach, and the real-time nature of responses. Traditional customer acquisition models mainly rely on manual screening methods, such as telephone marketing and offline promotion. These methods are not only costly in terms of manpower and inefficient in execution, but also limited by the experience and subjective judgment of sales personnel, resulting in generally low customer conversion rates. More importantly, these methods are difficult to scale and dynamically optimize, and cannot adapt to rapidly changing market demands, severely restricting the competitiveness of companies in emerging markets.

[0026] At the same time, companies generally face the problem of data silos in customer data management. There is a lack of effective integration and collaborative analysis mechanisms between internal knowledge bases and external publicly available internet data (such as bidding information, news updates, social media sentiment, and corporate change announcements). This fragmentation prevents companies from building comprehensive and dynamic customer profiles, resulting in the loss of numerous potential business opportunities. For example, even if a company is marked as a "potential buyer" in its internal system, it is easy to miss the most suitable opportunity to reach them if the bidding announcements it publishes on the procurement platform or the strategic adjustment information disclosed on its official website are not captured in time. In addition, existing customer tagging systems mostly adopt static and rule-based construction methods, relying on manually predefined tag dimensions (such as industry, size, and region), with long update cycles, making it difficult to reflect the dynamic changes in customer behavior preferences, business status, or procurement intentions in real time, leading to lagging marketing strategies or even misjudgments.

[0027] In the face of increasingly fierce market competition, the speed and accuracy of customer reach have evolved into a key barrier to a company's customer acquisition capabilities. Currently, the mainstream customer discovery solutions on the market still have significant shortcomings: Solution 1, the customer acquisition model based on the traditional tag system, is essentially a passive matching driven by static profiles. Tags rely on manual experience to define, lack self-learning and evolution capabilities, and cannot identify customers' implicit needs.

[0028] Solution 2, the customer acquisition method based on a single enterprise knowledge base, is limited to the mining of historical interaction data, has a narrow vision, is difficult to discover incremental customers or predict cross-industry opportunities, and lacks the ability to perceive the external market environment.

[0029] Solution 3, the technical path of using Internet search combined with keyword matching, although it can obtain massive amounts of external data, faces the problems of large data noise and serious information redundancy, and only stays at the surface text matching level, unable to understand unstructured data.The deep semantics and commercial intent in the text lead to a low effective clue extraction rate and a high false positive rate.

[0030] Therefore, the present invention aims to propose an intelligent customer discovery method and system that integrates multi-source heterogeneous data. It can integrate various customer acquisition methods through artificial intelligence, realize data fusion and collaborative analysis, reduce costs, improve the depth and breadth of customer acquisition, and thus improve the conversion rate.

[0031] Figure 1 is a schematic diagram of the customer acquisition system of the present invention. As shown in Figure 1, the customer acquisition system includes a planner 1, an Internet customer acquisition intelligent agent 2, a knowledge base customer acquisition intelligent agent 3, a tag customer acquisition intelligent agent 4, and a summary intelligent agent 5.

[0032] When a user needs to use the customer acquisition function, they input query information through the human-computer interaction interface provided by the customer acquisition system. The query information is the question the user needs to ask. For example: I want to find a batch of factories that can make knitted hats. After receiving the query information input by the user, the customer acquisition system executes the relevant customer acquisition process. The customer acquisition system and the human-computer interaction interface can be implemented on the same device. That is, the customer acquisition system is directly deployed on local devices such as mobile phones, laptops, or desktop computers. After the user inputs, the background expansion, sub-task decomposition and execution are completed locally immediately, and the results are returned to the interface in real time. Alternatively, it can be deployed in a split manner, that is, the human-computer interaction interface is located on terminals such as mobile phones, laptops, and desktop computers, and the customer acquisition system is placed on a server. The terminal and the server communicate through the network. The server completes the background expansion, sub-task decomposition and intelligent agent execution, and the results are sent back to the terminal, realizing centralized computing power and unified maintenance. The human-computer interaction interface can be flexibly selected according to the scenario, such as web pages, mini-programs, and Apps (Applications).

[0033] Further, the execution process of the customer acquisition system can be divided into three stages, namely the planning stage, the execution stage and the summary stage.

[0034] During the planning phase, after obtaining the user's query information, the planner 1 acquires the prior knowledge corresponding to the query information and generates subtasks based on the prior knowledge and the query information. These subtasks include internet customer acquisition subtasks, knowledge base customer acquisition subtasks, and tag customer acquisition subtasks. Specifically, user questions may contain the latest buzzwords, industry terms, abbreviations, and even memes. To avoid deviations in the planner's subtask generation process and the execution process of the independent customer acquisition agent due to a lack of prior knowledge, and to avoid selecting incorrect or missing tags, while also improving the reasoning ability of the large model, the planner calls the internet retrieval link before generating subtasks to retrieve, analyze, and extract background knowledge from the user's input query information, generating prior knowledge such as concept definitions, problem objectives, and analysis frameworks. Then, the planner encodes the prior knowledge and query information into a prompt through prompt engineering, and then uses the large model (CoT) to generate the subtasks.The Prompt (Chain-of-Thought Prompting) outputs the task flow. CoT Prompt engineering is a prompting engineering technique that guides large language models to perform more complex and logical reasoning. By adding examples of progressive reasoning to the prompt, the model is guided to mimic this reasoning process, thereby improving its performance on complex tasks.

[0035] Further, during the planning phase, a runtime state manager and a runtime state list are established simultaneously to detect the event streams returned by each execution process during subsequent execution.

[0036] The planner 1 can be implemented using an Agent (intelligent agent or proxy). An Agent is a software entity or program capable of perceiving its environment, making autonomous decisions, and taking actions to achieve a specific goal. An Agent typically includes functions such as perception, decision-making, action, and goal orientation. Among them, the Agent obtains input information through perception; it processes the perceived information and decides what to do next by using a certain mechanism (such as a rule system, machine learning model, or large language model) through decision-making; the Action function is that the Agent influences the environment through executors (such as sending messages, calling APIs, modifying files, etc.). Goal orientation means that the Agent's behavior is to achieve one or more preset goals, rather than random actions.

[0037] In the execution phase, each customer acquisition agent executes the corresponding sub-task to obtain intermediate customer acquisition results. The intermediate customer acquisition results are stored in OSS (Object Storage Service), and the intermediate customer acquisition results include Internet customer acquisition results, knowledge base customer acquisition results, tag customer acquisition results, and customer acquisition reasons, etc.

[0038] Specifically, the Internet customer acquisition agent 2 executes the Internet customer acquisition sub-task to generate Internet customer acquisition results, and the Internet customer acquisition results are Markdown (Lightweight Markup Language) customer acquisition reports. The Internet customer acquisition intelligence receives the above query information and prior knowledge, and actively mines potential customer leads through publicly available Internet data (such as social media, industry forums, corporate websites, recruitment platforms, news, etc.) to generate a Markdown customer acquisition report. Markdown is a lightweight markup language that allows documents to be written in an easy-to-read and easy-to-write plain text format and then converted into valid XHTML (or HTML) documents.

[0039] The knowledge base customer acquisition intelligence 3 executes the knowledge base customer acquisition sub-task to generate knowledge base customer acquisition results, which serve as search conditions. The knowledge base customer acquisition intelligence uses the enterprise's internal knowledge base to generate search conditions, providing a basis for subsequent outreach strategies.

[0040] The tag customer acquisition intelligence 4 executes the tag customer acquisition sub-task to generate tag customer acquisition results, which serve as search conditions.The customer acquisition results are DSL search statements. The tag-based customer acquisition agent generates DSL (Domain-Specific Language) search statements based on the enterprise tag system (such as industry, technology stack, business model, compliance requirements), and connects to internal or external databases for customer screening.

[0041] In the summary stage, the summary agent 5 generates auxiliary customer acquisition results based on the intermediate customer acquisition results and outputs them.

[0042] Specifically, the Markdown customer acquisition report is identified using a predetermined recognition model to extract enterprise information. In its specification, page 5 / 13 of CN 121526660 A, the predetermined recognition model is the NER (Named Entity Recognition) model, which is used to automatically identify entities with specific meanings from unstructured text and classify them into predefined categories. Among them, the extracted enterprise information can be enterprise name, enterprise ID (enterprise identifier), etc. Enterprise ID is a number or string code used to uniquely identify an enterprise in a specific system, platform, or database.

[0043] The matched paragraphs are obtained using the search conditions, and the enterprise information corresponding to the matched paragraphs is obtained. The knowledge base customer acquisition agent executes the knowledge base customer acquisition sub-task to generate search conditions, uses the search conditions to search the predetermined knowledge base, obtains the matched paragraphs, and then obtains the enterprise information corresponding to the matched paragraphs. Specifically, firstly, an enterprise database is established, which stores relevant information about the enterprise, such as enterprise name, address, year of establishment, scale, official website, enterprise ID, email, etc. Correspondingly, the generated DSL search statement is the logical expression corresponding to the user's query information. Then, the matching enterprises are searched according to the DSL search statement, and the corresponding enterprise information is output.

[0044] The corresponding enterprise information is retrieved using the DSL search statement, and the obtained enterprise information is summarized to obtain the auxiliary customer acquisition results. The summarizing agent summarizes the results obtained from the above Internet customer acquisition, tag customer acquisition and knowledge base customer acquisition, performs deduplication, sorting and other processing, and generates auxiliary customer acquisition results. The auxiliary customer acquisition results can be a file in a specified format, such as Excel. The auxiliary customer acquisition results are sent to the human-computer interaction interface for display. Thus, by providing a reliable relationship chain through the knowledge base, capturing hotspots through Internet data, and systematizing through the tag structure, multi-source data complementarity can be achieved.

[0045] The auxiliary lead generation results include text results, table results, and full results. Text results can be natural language summaries, such as telling the user how many leads were found. Table results contain information for each lead, such as company name, contact information, production capacity, and confidence score, facilitating filtering and follow-up. Full results are complete data packages, making it convenient for users to download.Therefore, users can quickly and accurately filter out enterprises that need further customer acquisition methods based on the auxiliary customer acquisition results, providing intelligent assistance for customer acquisition.

[0046] In some embodiments, the process variables of the customer acquisition intelligent agent and the summary intelligent agent during the execution process are passed down through context, slots, etc., which facilitates the connection of different customer acquisition intelligent agents, supports the passing down of all events of different customer acquisition intelligent agents and the summary intelligent agent, and users will perceive smooth event output, improving user experience.

[0047] This embodiment of the invention obtains the user's query information, obtains the prior knowledge corresponding to the query information, generates Internet customer acquisition sub-tasks, knowledge base customer acquisition sub-tasks and tag customer acquisition sub-tasks based on the prior knowledge and query information, calls the customer acquisition intelligent agents corresponding to each sub-task to generate intermediate customer acquisition results, and generates auxiliary customer acquisition results based on the intermediate customer acquisition results through the summary intelligent agent and outputs them. Therefore, multiple customer acquisition methods can be integrated through artificial intelligence to provide intelligent assistance for customer acquisition, realize data fusion and collaborative analysis, reduce costs, improve the depth and breadth of customer acquisition, and thus improve the conversion rate.

[0048] Figure 2 is a flowchart of the customer acquisition method of this embodiment of the invention. As shown in Figure 2, the customer acquisition method of this embodiment includes the following steps: Step S100: Obtain the user's query information.

[0049] In this embodiment, when the user needs to use the customer acquisition function, he / she inputs query information through the human-computer interaction interface provided by the customer acquisition system. The query information is the question the user needs to ask. For example: I want to find a batch of factories that can make knitted hats.

[0050] Step S200: Obtain the prior knowledge corresponding to the query information.

[0051] In this embodiment, since the query information input by the user may contain the latest popular words, industry terms, abbreviations, and even memes, in order to avoid deviations in the subsequent execution process due to the lack of prior knowledge, and to avoid selecting the wrong or missing correct tags, and at the same time improve the reasoning ability of the customer acquisition agent, it is necessary to supplement and expand the query information.

[0052] Specifically, Figure 3 is a flowchart of obtaining prior knowledge according to this embodiment. As shown in Figure 3, obtaining the prior knowledge corresponding to the query information includes the following steps: Step S210: Call the Internet retrieval link to obtain the background knowledge corresponding to the query information.

[0053] In this embodiment, after obtaining the user's query information, the query information is segmented into candidate words by a word segmenter, and a general search engine searches for background knowledge of each candidate word.

[0054] Step S220: Generate the prior knowledge based on the background knowledge. The prior knowledge includes one or more of concept definitions, problem objectives, and analysis frameworks.

[0055] In this embodiment, a predetermined Prompt can be used to allow the large model to summarize the query information and background knowledge.In summary, prior knowledge is output, which includes one or more of the following: concept definition, problem objective, and analysis framework.

[0056] The concept definition refers to the relevant definition of keywords in the query information. The problem objective is the user's real needs, which may include key indicators and constraints. The analysis framework is the basis for subsequent customer acquisition results retrieval, filtering, sorting, etc.

[0057] Step S300: Generate sub-tasks based on the prior knowledge and the query information.

[0058] In this embodiment, sub-tasks corresponding to each customer acquisition agent are generated based on the prior knowledge and the query information. As mentioned above, the customer acquisition agents include Internet customer acquisition agents, knowledge base customer acquisition agents, and tag customer acquisition agents. Correspondingly, the sub-tasks include Internet customer acquisition sub-tasks, knowledge base customer acquisition sub-tasks, and tag customer acquisition sub-tasks.

[0059] Each sub-task includes task information and execution tools. The task information of Internet customer acquisition sub-tasks, knowledge base customer acquisition sub-tasks, and tag customer acquisition sub-tasks can be the same or different. For example, the task information can be the above-mentioned query information and prior knowledge. The execution tool corresponds to the corresponding customer acquisition intelligent agent. For example, the execution tool for the Internet customer acquisition sub-task is the Internet customer acquisition intelligent agent, the execution tool for the knowledge base customer acquisition sub-task is the knowledge base customer acquisition intelligent agent, and the execution tool for the tag customer acquisition sub-task is the tag customer acquisition intelligent agent.

[0060] Wherein, the customer acquisition intelligent agent is an Agent (intelligent agent or agent), which is a software entity or program that can perceive its environment, make autonomous decisions, and take actions to achieve specific goals. An Agent usually includes functions such as perception, decision-making, action, and goal orientation. Among them, the Agent obtains input information through the perception function; through the decision-making function, it uses a certain mechanism (such as a rule system, machine learning model, large language model) to process the perceived information and decide what to do next; the action function is that the Agent influences the environment through executors (such as sending messages, calling APIs, modifying files, etc.). Goal orientation means that the Agent's behavior is to achieve one or more preset goals, rather than random actions.

[0061] Further, the process of the customer acquisition intelligent agent generating intermediate customer acquisition results includes one or more of the following: keyword expansion, query information analysis, background knowledge and analysis suggestions, and output result analysis.

[0062] Step S400: Call the customer acquisition intelligence agent corresponding to each subtask to generate intermediate customer acquisition results.

[0063] In this embodiment, the customer acquisition intelligence agent corresponding to each subtask is called to execute the corresponding subtask and generate the corresponding intermediate customer acquisition results.

[0064] Specifically, Figure 4 is a flowchart of obtaining intermediate customer acquisition results according to an embodiment of the present invention. As shown in Figure 4, calling the customer acquisition intelligence agent corresponding to each subtask to generate intermediate customer acquisition results includes the following steps:Step S410: The Internet customer acquisition intelligent agent executes the Internet customer acquisition sub-task to generate Internet customer acquisition results, which are Markdown customer acquisition reports.

[0065] In this embodiment, the Internet customer acquisition intelligent agent receives the above query information and prior knowledge, and actively mines potential customer leads through public Internet data (such as social media, industry forums, corporate websites, recruitment platforms, news, etc.) to generate Markdown customer acquisition reports. Markdown is a lightweight markup language that allows documents to be written in an easy-to-read and easy-to-write plain text format and then converted into valid XHTML (or HTML) documents.

[0066] Specifically, after receiving the above query information and prior knowledge, the Internet customer acquisition intelligent agent parses the query information and prior knowledge to generate specific search targets and keywords. The Internet customer acquisition intelligent agent actively accesses public Internet data sources through API (Application Programming Interface) or other network technologies, and performs intelligent search based on the parsed keywords and search targets to locate relevant enterprises and extract relevant leads. Finally, the Agent organizes the analyzed clues into structured content and converts them into a Markdown customer acquisition report.

[0067] Step S420: The knowledge base customer acquisition sub-task is executed by the knowledge base customer acquisition agent to generate knowledge base customer acquisition results, which are search conditions.

[0068] In this embodiment, the tag customer acquisition agent receives the user's query information and prior knowledge, and generates search conditions based on the query information and prior knowledge to provide a basis for subsequent outreach strategies.

[0069] The search conditions include one or more of the following combinations: keywords, phrases, Boolean logic (such as AND, OR), filters (such as industry type, geographical range), or semantic queries (such as natural language questions).

[0070] Step S430: The tag customer acquisition sub-task is executed by the tag customer acquisition agent to generate tag customer acquisition results, which are DSL search statements.

[0071] In this embodiment, the tag-based customer acquisition agent generates DSL (Domain-Specific Language) search statements based on the enterprise tag system (such as industry, technology stack, business model, compliance requirements), and connects to internal or external databases for customer screening.

[0072] Specifically, the tag-based customer acquisition agent receives the user's query information and prior knowledge, parses the query information, and identifies the user's core intent. Then, it extracts entities and condition values ​​from the query information, such as region, industry, and enterprise size. Combining relevant information from prior knowledge, the extracted entities are combined into query logic. For example, if the user's question involves multiple conditions...The document uses logical operators (AND, OR) to construct condition combinations. Based on the above analysis, the tag-based customer acquisition agent outputs a structured DSL search statement.

[0073] Step S500: The summarizing agent generates and outputs auxiliary customer acquisition results based on the intermediate customer acquisition results.

[0074] In this embodiment, the summarizing agent generates and outputs auxiliary customer acquisition results based on each intermediate customer acquisition result.

[0075] Figure 5 is a flowchart of generating auxiliary customer acquisition results according to an embodiment of the present invention. As shown in Figure 5, the process of generating and outputting auxiliary customer acquisition results based on the intermediate customer acquisition results by the summarizing agent includes the following steps: Step S510: The Markdown customer acquisition report is identified using a predetermined recognition model to extract enterprise information.

[0076] In this embodiment, the Markdown customer acquisition report is generated by the Internet customer acquisition sub-task executed by the Internet customer acquisition agent, and the Markdown customer acquisition report is identified using a predetermined recognition model to extract enterprise information. The predetermined recognition model is the NER (Named Entity Recognition) model, which is used to automatically identify entities with specific meanings from unstructured text and classify them into predefined categories. The extracted enterprise information can be enterprise name, enterprise ID (enterprise identifier), etc. The enterprise ID is a number or string code used to uniquely identify an enterprise in a specific system, platform, or database.

[0077] Specifically, after obtaining the Markdown customer acquisition report, the Markdown customer acquisition report can be preprocessed to optimize the recognition effect. For example, the Markdown format marks can be removed, and the report content can be restored to coherent paragraph text. If the text is very long, it can be logically divided into blocks to prevent the model from losing focus due to the excessive length of the text and to facilitate the subsequent association of the identified entities with specific enterprise entries. Then, the predetermined NER model is used to process the preprocessed text to extract enterprise information.

[0078] Step S520: Use the search conditions to obtain the hit paragraphs and obtain the enterprise information corresponding to the hit paragraphs.

[0079] In this embodiment, the knowledge base customer acquisition agent executes the knowledge base customer acquisition sub-task to generate search conditions, uses the search conditions to search in the predetermined knowledge base, obtains the matched paragraphs, and then obtains the enterprise information corresponding to the matched paragraph, specification page 8 / 13, 11 CN 121526660 A.

[0080] Specifically, firstly, a structured knowledge base is constructed to store text content related to the enterprise (such as product descriptions, industry reports, news articles, etc.), and these contents are organized into searchable paragraphs, wherein a paragraph can be a text fragment, such as a sentence, a paragraph, or a document fragment.

[0081] The process of building a knowledge base can be divided into stages such as data collection, data preprocessing, knowledge structuring, index building, and knowledge base maintenance. For data collection, raw data is obtained from various sources, such as publicly available company directories, industry websites, social media, internal documents, or third-party data providers. This data may include company profiles, product information, market dynamics, customer reviews, etc. For data preprocessing, the collected raw data is cleaned and standardized to remove noise (such as HTML tags and duplicate content), correct errors (such as spelling errors), standardize formats (such as dates and units), and ensure data consistency and quality. Preprocessing may also include natural language processing (NLP) operations such as text segmentation and stop word removal for subsequent indexing. For knowledge structuring, the processed data is organized into easily searchable units, i.e., paragraphs. Each paragraph typically corresponds to a semantically complete text block and is associated with a specific company entity (e.g., one paragraph might describe a company's core product, while another might relate to the company's market positioning). Simultaneously, the knowledge base establishes metadata indexes, such as adding tags (e.g., industry classification, geographic location, keywords) to each paragraph for rapid filtering and matching. For index building, in order to support fast retrieval, the knowledge base will build a search index (such as an inverted index or a vector index) to map the keywords or semantic features in the paragraphs to their positions, so that the relevant paragraphs can be quickly located during the query. If the knowledge base needs to support semantic search, it may also use embedded vectors to represent the meaning of the paragraphs, thereby achieving more intelligent matching. For knowledge base maintenance, the knowledge base needs to be updated regularly to reflect changes in enterprise information (such as the addition of new companies or the obsolescence of old data), such as version control, incremental updates and quality control, to ensure the accuracy and timeliness of the knowledge base.

[0082] After obtaining the search conditions, the search is used to search for and return matching text paragraphs in the knowledge base. These matched paragraphs are text fragments that are highly related to the search conditions and can be used to extract enterprise information later. Among them, the system applies the search conditions to the index structure of the knowledge base to perform matching calculations. The search for matching text paragraphs in the knowledge base according to the search conditions can be determined according to the design of the knowledge base. Among them, the search methods can be keyword matching and semantic matching, etc. Among them, keyword matching is to use an inverted index to quickly find paragraphs containing specific words. For example, the search term "cloud computing services" will return all paragraphs containing these words. Semantic matching converts the search terms into vector representations, then calculates their similarity (such as cosine similarity) to the paragraph vectors, returning the paragraphs that are semantically closest. In some embodiments, the search results can also be filtered and sorted, incorporating metadata filtering during the search process (such as retrieving only paragraphs from specific industries or regions), and sorting the results (by relevance, timestamps, etc.) to ensure that the most relevant paragraphs are returned first.

[0083] Further, after obtaining the matched paragraphs, the enterprise information corresponding to each paragraph is determined.

[0084] Step S530: Use the DSL retrieval statement to retrieve the corresponding enterprise information.

[0085] In this embodiment, the tag-based customer acquisition agent executes the tag-based customer acquisition subtask to generate the DSL retrieval statement, and uses the DSL retrieval statement to retrieve the corresponding enterprise information.

[0086] Specifically, firstly, an enterprise database is established, which stores relevant information about the enterprise, such as enterprise name, address, year of establishment, size, official website, enterprise ID, email, etc. Correspondingly, the generated DSL retrieval statement is the logical expression corresponding to the user's query information. Then, the matching enterprises are retrieved according to the DSL retrieval statement, and the corresponding enterprise information is output.

[0087] More specifically, the DSL retrieval statement is parsed, converted into a low-level query instruction (such as an SQL statement or API call), and the retrieval is executed to obtain a list of matching enterprises. Each record contains the field information specified in the DSL.

[0088] In some embodiments, the present invention can also parse Internet text in real time through a large model and update customer tags in the enterprise database.

[0089] That is, a heterogeneous customer acquisition agent is formed by combining multiple customer acquisition agents (Internet customer acquisition agent, knowledge base customer acquisition agent, and tag customer acquisition agent). This heterogeneous customer acquisition agent produces heterogeneous customer acquisition results. That is, the Internet customer acquisition agent will output a Markdown customer acquisition report, the tag customer acquisition agent will output a DSL search statement, and the knowledge base customer acquisition agent will output search conditions. Then, for the Internet customer acquisition results, the Markdown customer acquisition report is identified, and the enterprise name and enterprise ID are extracted using the enterprise name recognition NER model. For tag customer acquisition, the corresponding enterprise name and enterprise ID are retrieved using the DSL search statement. For knowledge base customer acquisition, the enterprise name and enterprise ID corresponding to the hit paragraph are retrieved.

[0090] Among them, tag customer acquisition is based on customer feature tags (such as industry, size, tax credit, enterprise operating indicators, etc.) for accurate screening and matching. Knowledge base customer acquisition utilizes unstructured text such as knowledge bases accumulated internally / externally by enterprises to acquire customers. Internet customer acquisition involves obtaining external internet enterprise information (such as official websites, social media, and bidding information) through the internet.

[0091] Step S540: Summarize the acquired enterprise information to obtain the auxiliary customer acquisition results.

[0092] In this embodiment, a summarizing agent is used to summarize the results obtained from the above-mentioned internet customer acquisition, tag customer acquisition, and knowledge base customer acquisition to obtain the auxiliary customer acquisition results.

[0093] Wherein, the summarizing agent summarizes the results obtained from the above-mentioned internet customer acquisition, tag customer acquisition, and knowledge base customer acquisition.In summary, after deduplication, sorting, and other processing, auxiliary customer acquisition results are generated. These results can be files in a specified format, such as Excel. The auxiliary customer acquisition results are then sent to the human-computer interaction interface for display. Thus, by providing a reliable relationship chain through the knowledge base, capturing hotspots through internet data, and systematizing the tag structure, multi-source data complementarity can be achieved. After obtaining the auxiliary customer acquisition results, users can quickly and accurately filter out enterprises that require further customer acquisition measures based on these results.

[0094] In some embodiments, the above customer acquisition system is compressed into a server-side microservice through model distillation to reduce computational costs.

[0095] In this embodiment of the invention, by obtaining the user's query information and the prior knowledge corresponding to the query information, internet customer acquisition sub-tasks, knowledge base customer acquisition sub-tasks, and tag customer acquisition sub-tasks are generated based on the prior knowledge and query information. The customer acquisition agents corresponding to each sub-task are called to generate intermediate customer acquisition results. The auxiliary customer acquisition results are generated and output by the summarizing agent based on the intermediate customer acquisition results. Therefore, artificial intelligence can be used to integrate various customer acquisition methods to provide intelligent assistance for customer acquisition, realize data fusion and collaborative analysis, reduce costs, improve the depth and breadth of customer acquisition, and thus improve the conversion rate.

[0096] In some embodiments, the customer acquisition method further includes: step S600, outputting the events of the customer acquisition agent and the summary agent.

[0097] In this embodiment, the process variables of the customer acquisition agent and the summary agent during the execution process are passed down through context, slots, etc., which facilitates the connection of different customer acquisition agents, supports the passing down of all events of different customer acquisition agents and the summary agent, and users will perceive smooth event output, improving user experience.

[0098] Figure 6 is a schematic diagram of the customer acquisition task execution process of an embodiment of the present invention. As shown in Figure 6, the customer acquisition task execution process can be divided into the following processes: In the planning stage, the overall task is planned to obtain the execution flow. For example, the execution flow includes tool calling, keywords and definitions, task splitting, task execution, summary sub-tasks, task completion, etc.

[0099] For tool invocation, the tool is used to invoke the customer acquisition agent required for this customer acquisition task, such as an internet customer acquisition agent, a knowledge base customer acquisition agent, and a tag customer acquisition agent.

[0100] For keywords and definitions, after obtaining the user's query information, the tool invokes the internet retrieval link to obtain the background knowledge corresponding to the query information, and generates the prior knowledge based on the background knowledge. The prior knowledge includes one or more of the following: definition, problem objective, and analysis framework.

[0101] For task decomposition, the customer acquisition task is divided into multiple sub-tasks, including an internet customer acquisition sub-task, a knowledge base customer acquisition sub-task, and a tag customer acquisition sub-task.

[0102] For task execution, the customer acquisition agent executes the corresponding sub-tasks to obtain intermediate customer acquisition results. In the embodiment shown in Figure 6, taking knowledge base customer acquisition as an example, when the knowledge base customer acquisition agent executes the corresponding sub-tasks, it is divided into processes such as keyword expansion, problem splitting, background knowledge and analysis suggestions, and output result analysis. Among them, keyword expansion is to convert words in prior knowledge into search fields of the platform; problem splitting is to selectively split the query information input by the user into one or more problem targets according to the complexity of the problem. Background knowledge analysis and suggestions are to analyze the above-mentioned prior knowledge, list relevant knowledge, and give subsequent suggestions. Output result analysis refers to giving relevant suggestions for the intermediate customer acquisition results output this time, such as the customer acquisition reasons corresponding to each enterprise.

[0103] For the summary sub-task, the summary agent summarizes the results obtained from the above-mentioned Internet customer acquisition, tag customer acquisition and knowledge base customer acquisition, performs deduplication, sorting and other processing, and generates auxiliary customer acquisition results. The auxiliary customer acquisition results can be a file in a specified format, such as Excel. The auxiliary customer acquisition results are sent to the human-computer interaction interface for display.

[0104] Upon task completion, the user is prompted that the task execution is complete.

[0105] Figure 7 is a schematic diagram of the human-computer interaction interface of an embodiment of the present invention. Figure 7 shows the display interface for assisting customer acquisition results, including a dialogue area A1 and a result display area A3.

[0106] The dialogue area A1 is used to display the user's historical dialogues and includes an input box A2, which can receive new dialogues from the user.

[0107] The result display area A3 is used to display the assisting customer acquisition results and includes a download control A4 and a result display area A5. The user can click the download control to download the results to the local storage space.

[0108] In summary, since the query information output by the user may contain the latest buzzwords, industry terms, abbreviations, and even memes, in order to avoid the planner and customer acquisition agent from experiencing illusions due to a lack of prior knowledge during execution, to avoid selecting the wrong or missing labels, and to improve the reasoning ability of the large model, the customer acquisition system will call the Internet retrieval link before execution parsing to retrieve, analyze, and extract background knowledge from the user's query information, generate prior knowledge such as concept definitions, problem objectives, and analysis frameworks, and improve the recognition results of subsequent steps. Meanwhile, the customer acquisition system provides dynamic planning capabilities, breaking down user problems into multiple sub-problems based on their complexity, thus expanding customer acquisition coverage. Furthermore, the system provides agent execution environments and process persistence capabilities for different customer acquisition agents and summarizing agents, enabling unified scheduling and management. It supports serial execution, parallel execution, and cached execution (parallel execution, background output), with process variables passed down through context and slots for easy connection between different customer acquisition agents. Additionally, the system offers a rich set of system and process events, supporting…By transmitting different customer acquisition agents and summarizing all events of the agents, users will perceive a smooth event output, improving the user experience.

[0109] In this embodiment of the invention, by obtaining the user's query information and the prior knowledge corresponding to the query information, the invention generates Internet customer acquisition sub-tasks, knowledge base customer acquisition sub-tasks, and tag customer acquisition sub-tasks based on the prior knowledge and query information. The invention then calls the customer acquisition agents corresponding to each sub-task to generate intermediate customer acquisition results. The summary agent generates and outputs auxiliary customer acquisition results based on the intermediate customer acquisition results. Thus, multiple customer acquisition methods can be integrated through artificial intelligence to provide intelligent assistance for customer acquisition, realize data fusion and collaborative analysis, reduce costs, improve the depth and breadth of customer acquisition, and thereby improve the conversion rate.

[0110] Figure 8 is a schematic diagram of the customer acquisition device of this embodiment of the invention. As shown in Figure 8, the customer acquisition device of this embodiment of the invention includes a query information acquisition unit 81, a prior knowledge acquisition unit 82, a sub-task generation unit 83, an intermediate customer acquisition result generation unit 84, and an auxiliary customer acquisition result generation unit 85. Among them, the query information acquisition unit 81 is used to acquire the user's query information. The prior knowledge acquisition unit 82 is used to acquire the prior knowledge corresponding to the query information. The subtask generation unit 83 is used to generate subtasks based on the prior knowledge and the query information. The subtasks include Internet customer acquisition subtasks, knowledge base customer acquisition subtasks, and tag customer acquisition subtasks. The intermediate customer acquisition result generation unit 84 is used to call the customer acquisition intelligent agent corresponding to each subtask to generate intermediate customer acquisition results. The auxiliary customer acquisition result generation unit 85 is used to generate auxiliary customer acquisition results based on the intermediate customer acquisition results through the summarizing intelligent agent and output them.

[0111] In this embodiment of the invention, by acquiring the user's query information, acquiring the prior knowledge corresponding to the query information, generating Internet customer acquisition subtasks, knowledge base customer acquisition subtasks, and tag customer acquisition subtasks based on the prior knowledge and query information, calling the customer acquisition intelligent agent corresponding to each subtask to generate intermediate customer acquisition results, and generating auxiliary customer acquisition results based on the intermediate customer acquisition results through the summarizing intelligent agent and outputting them. Therefore, artificial intelligence can be used to integrate various customer acquisition methods to provide intelligent assistance for customer acquisition, realize data fusion and collaborative analysis, reduce costs, improve the depth and breadth of customer acquisition, and thus improve the conversion rate.

[0112] Figure 9 is a schematic diagram of an electronic device according to an embodiment of the present invention. In this embodiment, the electronic device 9 includes a server, a terminal, etc. As shown in Figure 9, the electronic device 9 includes at least one processor 91; and a memory 92 communicatively connected to at least one processor 91; and a communication component 93 communicatively connected to a scanning device, wherein the communication component 93 receives and sends data under the control of the processor 91; wherein the memory 92 stores instructions that can be executed by at least one processor 91, and the instructions are executed by at least one processor 91 to implement the above-mentioned data processing method for assisting customer acquisition.

[0113] Specifically, the electronic device includes one or more processors 91 and a memory 92, with one processor 91 as an example in FIG. 9. The processor 91 and the memory 92 can be connected via a bus or other means, with a bus connection as an example in FIG. 9. The memory 92, as a non-volatile computer-readable storage medium, can be used to store non-volatile software programs, non-volatile computer-executable programs, and modules. The processor 91 executes various functional applications and data processing of the device by running the non-volatile software programs, instructions, and modules stored in the memory 92, thereby realizing the above-mentioned data processing method for assisting customer acquisition.

[0114] The memory 92 may include a program storage area and a data storage area, wherein the program storage area may store the operating system and at least one application program required for a function; the data storage area may store an option list, etc. In addition, the memory 92 may include high-speed random access memory and may also include non-volatile memory, such as at least one disk storage device, flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 92 may optionally include memory remotely located relative to the processor 91, and these remote memories can be connected to external devices via a network. Examples of the aforementioned networks include, but are not limited to, the Internet, enterprise intranets, local area networks, mobile communication networks, and combinations thereof.

[0115] One or more modules are stored in the memory 92, and when executed by one or more processors 91, they execute the data processing method for assisting customer acquisition in any of the above method embodiments.

[0116] The above-mentioned product can execute the method provided in the embodiments of this application, and has the corresponding functional modules and beneficial effects of the method execution. For technical details not described in detail in this embodiment, please refer to the method provided in the embodiments of this application.

[0117] The embodiments of the present invention obtain user query information, obtain prior knowledge corresponding to the query information, generate Internet customer acquisition sub-tasks, knowledge base customer acquisition sub-tasks, and tag customer acquisition sub-tasks based on prior knowledge and query information, respectively call the customer acquisition intelligent agent corresponding to each sub-task to generate intermediate customer acquisition results, and generate and output auxiliary customer acquisition results based on the intermediate customer acquisition results through the summarizing intelligent agent. Thus, multiple customer acquisition methods can be integrated through artificial intelligence to provide intelligent assistance for customer acquisition, realize data fusion and collaborative analysis, reduce costs, improve the depth and breadth of customer acquisition, and thus improve the conversion rate.

[0118] Another embodiment of the present invention relates to a non-volatile storage medium for storing a computer-readable program, the computer-readable program being used by a computer to execute some or all of the above-described method embodiments.

[0119] That is, those skilled in the art will understand that implementing all or part of the steps in the methods of the above embodiments is possible. (See page 12 / 13 of the specification, 15 CN 121526660 A)The process is accomplished by instructing related hardware through a program. This program is stored in a storage medium and includes several instructions to cause a device (which may be a microcontroller, chip, etc.) or processor to execute all or part of the steps of the methods described in the embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0120] The above descriptions are merely preferred embodiments of this application and are not intended to limit this application. For those skilled in the art, this application can have various modifications and variations. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of this application should be included within the protection scope of this application. Instruction Manual 13 / 13 Page 16 CN 121526660 A Figure 1 Instruction Manual Figure 1 / 8 Page 17 CN 121526660 A Figure 2 Instruction Manual Figure 2 / 8 Page 18 CN 121526660 A Figure 3 Figure 4 Instruction Manual Figure 3 / 8 Page 19 CN 121526660 A Figure 5 Instruction Manual Figure 4 / 8 Page 20 CN 121526660 A Figure 6 Instruction Manual Figure 5 / 8 Page 21 CN 121526660 A Figure 7 Instruction Manual Figure 6 / 8 Page 22 CN 121526660 A Figure 8 Instruction Manual Figure 7 / 8 Page 23 CN 121526660 A Figure 9 Instruction Manual Figure 8 / 8 Page 24 CN 121526660 A Abstract Embodiments of the present application disclose a data processing method, system and device for assisting in customer acquisition. The method comprises obtaining query information of a user, obtaining prior knowledge corresponding to the query information, generating an internet customer acquisition subtask, a knowledge base customer acquisitionsubtask and a label customer acquisition subtask according to the prior knowledge and the query information, respectively invoking a customer acquisition agent corresponding to each subtask to generate an intermediate customer acquisition result, and generating and outputting an auxiliary customer acquisition result according to the intermediate customer acquisition result. Thus, the artificial intelligence can integrate various customer acquisition methods to provide intelligent assistance for customer acquisition, realize data fusion and collaborative analysis, reduce costs, improve the depth and breadth of customer acquisition, and further improve the conversion rate.

Claims

1. A data processing method to assist in customer acquisition, characterized in that, The method includes: Obtain the user's query information; Obtain the prior knowledge corresponding to the query information; Subtasks are generated based on the prior knowledge and the query information, including internet customer acquisition subtasks, knowledge base customer acquisition subtasks, and tag-based customer acquisition subtasks; Each subtask's corresponding customer acquisition agent is invoked to generate intermediate customer acquisition results; The intelligent agent generates and outputs auxiliary customer acquisition results based on the intermediate customer acquisition results.

2. The method according to claim 1, characterized in that, The prior knowledge corresponding to the query information includes: The internet retrieval process is invoked to obtain the background knowledge corresponding to the query information; The prior knowledge is generated based on the background knowledge, and the prior knowledge includes one or more of concept definitions, problem objectives, and analysis frameworks.

3. The method according to claim 1, characterized in that, The customer acquisition intelligent agent generates intermediate customer acquisition results by including one or more of the following: keyword expansion, query information analysis, background knowledge and analysis suggestions, and output result analysis.

4. The method according to claim 1, characterized in that, The customer acquisition intelligent agent includes an internet customer acquisition intelligent agent, a knowledge base customer acquisition intelligent agent, and a tag-based customer acquisition intelligent agent; The step of calling the customer acquisition agent corresponding to each subtask to generate intermediate customer acquisition results includes: The internet customer acquisition intelligent agent executes the internet customer acquisition sub-tasks to generate internet customer acquisition results, which are Markdown customer acquisition reports. The knowledge base customer acquisition agent executes the knowledge base customer acquisition sub-task to generate knowledge base customer acquisition results, which serve as search criteria. The tag-based customer acquisition agent executes the tag-based customer acquisition sub-task to generate tag-based customer acquisition results, which are DSL search statements.

5. The method according to claim 4, characterized in that, The process of summarizing and generating auxiliary customer acquisition results based on the intermediate customer acquisition results and outputting them includes: The Markdown customer acquisition report is identified using a predefined recognition model to extract enterprise information; Use the search criteria to obtain the matched paragraphs, and obtain the enterprise information corresponding to the matched paragraphs; Use the DSL search statement to retrieve the corresponding enterprise information; The acquired enterprise information is summarized to obtain the aforementioned customer acquisition assistance results.

6. The method according to any one of claims 1-5, characterized in that, The method further includes: Output the events of the customer acquisition agent and the summary agent.

7. A data processing system for assisting in customer acquisition, characterized in that, The system includes: The planner is used to obtain prior knowledge corresponding to the query information after obtaining the user's query information, and generate sub-tasks based on the prior knowledge and the query information. The sub-tasks include Internet customer acquisition sub-tasks, knowledge base customer acquisition sub-tasks, and tag customer acquisition sub-tasks. Multiple customer acquisition agents are used to execute corresponding sub-tasks to obtain intermediate customer acquisition results; The intelligent agent is used to generate and output auxiliary customer acquisition results based on the intermediate customer acquisition results.

8. The system according to claim 7, characterized in that, The planner is used to invoke Internet retrieval links to obtain background knowledge corresponding to the query information, and generate prior knowledge based on the background knowledge. The prior knowledge includes one or more of concept definitions, problem objectives, and analysis frameworks.

9. The system according to claim 7, characterized in that, The customer acquisition intelligent agent generates intermediate customer acquisition results by including one or more of the following: keyword expansion, query information analysis, background knowledge and analysis suggestions, and output result analysis.

10. The system according to claim 9, characterized in that, The customer acquisition intelligence agent includes an internet customer acquisition intelligence agent, a knowledge base customer acquisition intelligence agent, and a tag customer acquisition intelligence agent. The internet customer acquisition intelligence agent executes the internet customer acquisition sub-tasks to generate internet customer acquisition results, which are Markdown customer acquisition reports. The knowledge base customer acquisition intelligence agent executes the knowledge base customer acquisition sub-tasks to generate knowledge base customer acquisition results, which are search criteria. The tag-based customer acquisition agent executes the tag-based customer acquisition sub-task to generate tag-based customer acquisition results, which are DSL search statements.

11. The system according to claim 10, characterized in that, The summary agent is used to identify the Markdown customer acquisition report using a predetermined recognition model to extract enterprise information; it uses the search conditions to obtain the matched paragraphs and obtains the enterprise information corresponding to the matched paragraphs. Use the DSL search statement to retrieve the corresponding enterprise information; summarize the obtained enterprise information to obtain the auxiliary customer acquisition results.

12. A data processing device for assisting in customer acquisition, characterized in that, The device includes: The query information retrieval unit is used to retrieve the user's query information; A prior knowledge acquisition unit is used to acquire prior knowledge corresponding to the query information. The subtask generation unit is used to generate subtasks based on the prior knowledge and the query information. The subtasks include internet customer acquisition subtasks, knowledge base customer acquisition subtasks, and tag customer acquisition subtasks. The intermediate customer acquisition result generation unit is used to call the customer acquisition intelligent agents corresponding to each subtask to generate intermediate customer acquisition results. The auxiliary customer acquisition result generation unit is used to generate and output auxiliary customer acquisition results based on the intermediate customer acquisition results by summarizing the intelligent agent.

13. An electronic device comprising a memory and a processor, characterized in that, The memory is used to store one or more computer program instructions, wherein the one or more computer program instructions are executed by the processor to implement the method as described in any one of claims 1-6.

14. A computer-readable storage medium storing computer program instructions thereon, characterized in that, The computer program instructions, when executed by a processor, implement the method as described in any one of claims 1-6.

Images