Customer query method and device, medium and product

Abstract

The invention provides a customer query method and device, a medium and a product, relates to the technical field of information query, and is used for accurately querying potential customers meeting conditions when business personnel query potential customers. The method comprises the steps of obtaining a to-be-queried target description text; the target description text is used for indicating standardized description about potential customers; determining at least one target keyword based on the target description text and a preset prompt word template; the target keyword is used for indicating semantic information matched with the potential customer characteristics; based on the at least one target keyword, obtaining an initial customer list; the initial customer list is used for indicating customer information queried in a local customer database; performing quality screening on the initial customer list to obtain a target customer list required by the user; the quality screening is used for indicating screening of customer information in the initial customer list in combination with unstructured semantics in the target description text, so that accurate query of potential customers is realized.

Description

Technical Field

[0001] This application relates to the field of information retrieval technology, and in particular to a customer retrieval method, device, medium and product. Background Technology

[0002] When conducting marketing, sales personnel first need to identify potential customers who meet the marketing criteria from among the numerous government and enterprise clients in the market.

[0003] When business personnel search for potential customers, the relevant solutions are usually based on natural language models to directly query relevant information about potential customers. However, solutions based on natural language models cannot recognize ambiguous semantics and complex business concepts, and therefore cannot accurately retrieve potential customers who meet the criteria. Summary of the Invention

[0004] This application provides a customer query method, device, medium, and product, which enables business personnel to accurately query potential customers who meet the criteria when querying potential customers.

[0005] Firstly, this application provides a customer inquiry method, including: Retrieve the target description text to be queried; the target description text is used to indicate a standardized description of a potential customer. Based on the target description text and the preset prompt word template, at least one target keyword is determined; the target keyword is used to indicate semantic information that matches the characteristics of potential customers; Based on at least one target keyword, obtain an initial customer list; the initial customer list is used to indicate customer information retrieved from the local customer database. The initial customer list is subjected to quality screening to obtain the target customer list required by the user; the quality screening is used to instruct the filtering of customer information in the initial customer list in combination with the unstructured semantics in the target description text.

[0006] The technical solution provided in this application offers at least the following beneficial effects: By using a target description text that indicates a standardized description of potential customers and a preset prompt word template, at least one target keyword is determined to indicate semantic information matching the characteristics of potential customers. This allows for accurate identification of fuzzy semantics in the natural language input by the user through each target keyword. Based on at least one target keyword, an initial customer list indicating customer information retrieved from a local customer database is obtained. This initial customer list is then subjected to quality screening to obtain the target customer list required by the user. Furthermore, by combining the unstructured semantics in the target description text with the initial customer list for quality screening, the filtering and verification of customer information in the initial customer list is achieved. Thus, through accurate identification of the semantics of the natural language input by the user and the filtering and verification of customer information in the initial customer list, potential customers can be accurately retrieved.

[0007] One possible implementation involves obtaining the target description text to be queried, including: Obtain the data query request input by the user; the data query request is used to indicate a vague description about potential customers; Upon successful authorization verification of the data query request, the data query request is converted into target description text using a preset first prompt word template. Among them, the permission verification is used to indicate whether the geographical restriction keywords in the data query request meet the geographical range that the user can query, and the first prompt word template is used to transform the data query request from a vague description into a standardized description.

[0008] Another possible implementation involves converting the data query request into target descriptive text using a pre-defined first prompt word template, including: Retrieve structured standard keywords from data query requests. Standard keywords are used to indicate the names of benchmark customers similar to potential customers, descriptions of the industries in which potential customers operate, or the names of potential customers in data query requests. Fill the fields to be filled in with the standard keywords into the first prompt word template to obtain the first prompt word; the first prompt word is used to indicate how to convert the standard keywords into a standard text description of the potential customer's industry and core business. Input the first prompt word into the preset large language model to obtain the target description text output by the large language model.

[0009] Another possible implementation involves determining at least one target keyword based on the target description text and a pre-defined prompt template, including: Identify the customer type in the target description text, which indicates whether the potential customer is a public sector entity or a private sector entity; In response to a customer type of "unit in the public sector", a preset second prompt word template is invoked to obtain the target keywords about the potential customer's name. The second prompt word template is used to determine the full name of the potential customer. In response to a client being a private sector entity, at least one preset prompt word template is invoked to obtain at least one target keyword about the potential client. The at least one target keyword includes a target keyword about the potential client's name, a target keyword about the potential client's industry, and a target keyword about the potential client's business scope.

[0010] Another possible implementation involves calling a preset second prompt word template to obtain at least one target keyword related to the customer's name, including: Populate the target description text into the fields to be filled in the second prompt word template to generate the second prompt word; the second prompt word is used to indicate the generation of the full name of the customer that matches the description of the target description text. The second prompt word is input into the preset large language model to obtain at least one keyword related to the customer's name output by the large language model, and the keyword with the highest matching degree related to the customer's name is taken as the corresponding target keyword.

[0011] Another possible implementation involves calling at least one preset prompt word template to obtain at least one target keyword about potential customers, including: The target description text is filled into the corresponding fields to be filled in the second prompt word template, the preset third prompt word template, and the fourth prompt word template, respectively, to generate the corresponding second prompt word, third prompt word, and fourth prompt word; The third prompt word template is used to determine the industry of the potential customer, and the fourth prompt word template is used to determine the business scope of the potential customer. The second prompt word is used to indicate the generation of the complete name of the customer that matches the target description text; the third prompt word is used to indicate the generation of the industry that matches the target description text; and the fourth prompt word is used to indicate the generation of the business scope that matches the target description text. Input the second, third, and fourth prompt words into the preset large language model to obtain at least one keyword about the customer's name, at least one keyword about the industry, and at least one keyword about the business scope output by the large language model. Among at least one keyword related to the customer's name, the keyword with the highest matching degree is determined as the target keyword related to the potential customer's name; among at least one keyword related to the industry, the keyword with the highest matching degree is determined as the target keyword related to the potential customer's industry; among at least one keyword related to the business scope, the keyword with the highest matching degree is determined as the target keyword related to the potential customer's business scope.

[0012] Another possible approach is to obtain an initial customer list based on at least one target keyword, including: Generate a target query statement based on at least one target keyword; the target query statement is a structured query statement. The system retrieves matching customer information from the local customer database using a target query statement. The local customer database pre-stores multiple customer information entries, including customer name, industry, and business scope. An initial customer list is generated based on the matched customer information.

[0013] Another possible implementation involves quality screening of the initial customer list to obtain the target customer list that the user needs to query, including: Based on the target description text, obtain the first requirement description and the second requirement description of unstructured semantics; the first requirement description is used to indicate the regular filtering conditions set by the user, and the second requirement description is used to indicate the non-regular filtering conditions set by the user. The initial customer list is screened for quality based on the first and second requirements descriptions to obtain the target customer list.

[0014] Another possible implementation involves performing quality screening on the initial customer list based on the first and second requirement descriptions to obtain a target customer list, including: Based on the first requirement description, the initial customer list is initially screened to obtain a preliminary customer list; the preliminary screening is used to indicate the filtering of customer information that does not meet the regular screening criteria. Based on the description of the second requirement, a search engine query instruction is generated, and the public information of each customer in the initial screening customer list is retrieved online through the search engine query instruction. Based on publicly available information from each customer, the initial customer list is validated and filtered to obtain the target customer list; the validation and filtering are used to indicate the filtering of customer information that does not meet the non-standard filtering conditions.

[0015] In a second aspect, this application provides an electronic device comprising: a processor and a memory; the memory storing processor-executable instructions; when the processor is configured to execute the instructions, causing the electronic device to implement the method of the first aspect described above.

[0016] Thirdly, this application provides a computer-readable storage medium comprising: computer software instructions; when the computer software instructions are executed in an electronic device, they cause the electronic device to implement the method described in the first aspect.

[0017] Fourthly, this application provides a computer program product comprising a computer program; when the computer program is run in an electronic device, the electronic device performs the method described in the first aspect.

[0018] The beneficial effects of the second to fourth aspects mentioned above are described in the corresponding description of the first aspect and will not be repeated here. Attached Figure Description

[0019] Figure 1 This application provides a schematic diagram of the architecture of a customer query system. Figure 2 A flowchart illustrating a customer query method provided in an embodiment of this application; Figure 3 A flowchart illustrating another customer query method provided in this application embodiment; Figure 4 A flowchart illustrating yet another customer query method provided in this application embodiment; Figure 5 A schematic diagram of a prompt word template provided in an embodiment of this application; Figure 6 A schematic diagram of another prompt word template provided in an embodiment of this application; Figure 7 A schematic diagram illustrating yet another prompt word template provided in an embodiment of this application; Figure 8 A flowchart illustrating yet another customer query method provided in this application embodiment; Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0020] In this article, the term "and / or" is merely a description of the relationship between related objects, indicating that there can be three relationships. For example, A and / or B can represent three situations: A exists alone, A and B exist simultaneously, and B exists alone.

[0021] To facilitate a clear description of the technical solutions of the embodiments of this application, the terms "first" and "second" are used in the embodiments of this application to distinguish the same or similar items with essentially the same function and effect. Those skilled in the art can understand that the terms "first" and "second" are not intended to limit the quantity or execution order.

[0022] Furthermore, the terms "comprising" and "having," and any variations thereof, used in the description of this application are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the steps or units listed, but may optionally include other steps or units not listed, or may optionally include other steps or units inherent to such process, method, product, or apparatus.

[0023] It should be noted that in the embodiments of this application, the words "exemplary" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design scheme described as "exemplary" or "for example" in the embodiments of this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of the words "exemplary" or "for example" is intended to present the relevant concepts in a specific manner.

[0024] In the description of this application, unless otherwise stated, "a plurality of" means two or more.

[0025] When business personnel query information about potential customers using natural language models (NLPs), while the NLPs in relevant solutions incorporate feature engineering or knowledge graphs, they fail to deeply integrate specific knowledge from the government and enterprise sectors, such as policy terminology, industry classification standards, and complex enterprise concepts. This results in the NLPs' inaccurate understanding of vague query intents rich in domain knowledge, such as "high-energy-consuming enterprises" or "similar to Company X," creating a semantic gap.

[0026] When using fixed feature libraries and fixed rule templates in feature engineering for retrieval and querying, the relevant solutions cannot support multi-dimensional and comprehensive query needs, such as querying based on marketing conditions such as industry, registered capital, region, and business status. As a result, the final output list of potential customers cannot meet the needs of marketers, and thus cannot accurately find potential customers who meet the criteria.

[0027] Furthermore, when business personnel query information about potential customers based on the natural language model, they also need to convert the user's input natural language into a corresponding SQL (Structured Query Language) query. The natural language model then uses this SQL query to retrieve relevant potential customers. However, when converting user-input natural language into SQL queries, the existing methods often fail to extract the unstructured semantics from the natural language, which may contain complex unstructured semantics describing business conditions. This results in an inaccurate list of potential customers obtained based on the query.

[0028] To address the aforementioned technical problems, this application provides a customer query method, device, medium, and product. Based on target description text indicating standardized descriptions of potential customers and preset prompt word templates, at least one target keyword is determined to indicate semantic information matching the characteristics of potential customers. This allows for accurate identification of fuzzy semantics in the natural language input by the user through each target keyword. Based on at least one target keyword, an initial customer list indicating customer information retrieved from a local customer database is obtained. This initial customer list undergoes quality screening to obtain a target customer list required by the user. Furthermore, by combining unstructured semantics in the target description text with the initial customer list for quality screening, the customer information in the initial customer list is verified. Thus, through accurate identification of the semantics of the natural language input by the user and verification of the customer information in the initial customer list, potential customers can be accurately retrieved.

[0029] The embodiments provided in this application will now be described in detail with reference to the accompanying drawings.

[0030] Figure 1 This is a schematic diagram of the architecture of a customer query system provided in an embodiment of this application. See also... Figure 1 The query system comprises a user interaction layer, a core processing layer, and a data layer. Data flows between these three layers.

[0031] In some embodiments, the user interaction layer obtains data query requests input by business personnel through an interactive interface, wherein the data query requests are used to indicate a vague description of potential customers.

[0032] The data layer comprises a local customer database, an external data source, and a domain knowledge database, all of which are accessed by the core processing layer. The local customer database contains pre-stored customer information; the external data source indicates networked external data interfaces; and the domain knowledge database provides descriptions of professional concepts from different industries. The core processing layer is implemented based on the corresponding processor.

[0033] In some embodiments, the core processing layer includes an intent recognition unit, a semantic query unit, and a quality verification unit. The intent recognition unit, semantic query unit, and quality verification unit sequentially process the query request to generate the target customer list required by the user.

[0034] The intent recognition unit converts the data query request sent by the user interaction layer into a standardized target description text by calling the domain knowledge database in the data layer. The semantic query unit determines at least one target keyword based on the target description text and the preset prompt word template, and generates a corresponding structured query statement based on the at least one target keyword to obtain an initial customer list. The quality verification unit performs quality screening on the initial customer list by calling an external data source to obtain the target customer list required by the user.

[0035] In some embodiments, after generating the target customer list needed by the user, the core processing layer returns the target customer list to the user interaction layer for display, so that business personnel can obtain information about potential customers from the target customer list.

[0036] It should be noted that the system architecture described in the embodiments of this application is for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and does not constitute a limitation on the technical solutions provided in the embodiments of this application. As those skilled in the art will know, with the evolution of system architecture, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0037] Figure 2 This is a flowchart illustrating a customer query method provided in an embodiment of this application. (In conjunction with...) Figure 2 As shown, the customer query method provided in this application is applied to Figure 1 The processor in the core processing layer. Combined with... Figure 2 As shown, this query method includes the following steps: S201. Obtain the target description text to be queried.

[0038] In some embodiments, query requests input by business personnel are obtained through a user interaction layer. Intent filtering is performed on these queries to filter out non-data-related queries and retain only data-related queries. Geographical access permissions are then applied to data-related queries. Query requests that pass permission checks are then standardized to obtain target description text.

[0039] The target description text is used to indicate a standardized description of a potential customer.

[0040] S202. Based on the target description text and the preset prompt word template, determine at least one target keyword.

[0041] In some embodiments, after obtaining the target description text, different retrieval modes are invoked according to the type of potential customer in the target description text, and the corresponding target keywords are obtained through the corresponding retrieval modes. The target keywords are used to indicate semantic information that matches the characteristics of potential customers.

[0042] For example, when the target description text indicates that the potential customer type is a public sector organization, the first search mode is invoked to retrieve the corresponding target keywords based on the target description text. The first search mode indicates a fuzzy search based on name.

[0043] For example, when the target description text identifies the potential customer as a private sector entity, the second search mode is invoked to retrieve matching target keywords based on the target description text. The second search mode indicates a method for combining searches based on multiple feature dimensions such as name and business scope.

[0044] S203. Obtain an initial customer list based on at least one target keyword.

[0045] In some embodiments, a target query statement is generated based on at least one target keyword. The target query statement is a structured query language (SQL).

[0046] In some embodiments, a target query statement is used to query matching customer information in a local customer database. The local customer database pre-stores multiple customer information items, including customer name, customer industry, and customer business scope. Based on the matching customer information, an initial customer list is generated, which is used to indicate the customer information obtained from the query in the local customer database.

[0047] S204. Perform quality screening on the initial customer list to obtain the target customer list required by the user.

[0048] In some embodiments, quality screening is used to instruct the filtering of customer information in an initial customer list based on unstructured semantics in the target description text. Here, unstructured semantics in the target description text are used to indicate data that does not have a predefined, fixed format, such as "no operational risk."

[0049] In this embodiment, at least one target keyword is determined by using a target description text that indicates a standardized description of potential customers and a preset prompt word template to indicate semantic information that matches the characteristics of potential customers. Based on the at least one target keyword, an initial customer list indicating customer information retrieved from a local customer database is obtained. The initial customer list is then quality-screened to obtain the target customer list required by the user. Thus, by combining the target keywords and quality-screening the initial customer list, the accuracy of potential customer queries is improved.

[0050] Figure 3This is a flowchart illustrating another customer query method provided in an embodiment of this application. (In conjunction with...) Figure 3 As shown, obtaining the target description text to be queried in step S201 above can be achieved through the following steps: S2011, Obtain the data query request input by the user.

[0051] In some embodiments, a data query request is used to instruct a user to enter a vague description of a potential customer.

[0052] In other embodiments, if the user inputs a non-data query request, such as casual conversation, the query process is terminated directly and an instruction prompt is returned to remind the user to input a data query request, thereby effectively saving system resources.

[0053] S2012. In response to the successful authorization verification of the data query request, the data query request is converted into target description text using the preset first prompt word template.

[0054] In some embodiments, after receiving a data query request, permission verification is performed on the data query request. The permission verification is used to indicate whether the geographical restriction keywords in the data query request meet the geographical range that the user can query. The first prompt word template is used to transform the data query request from a vague description into a standardized description.

[0055] For example, the system can accurately identify geographical restriction keywords (such as "nationwide", "Shandong Province", "Qingdao City") from the data query request, and compare the identified geographical restriction keywords with the user's preset permission geographical regions in terms of logical relationship (such as contain, included, intersecting, not intersecting, and cannot be determined).

[0056] If the region corresponding to the regional restriction keyword exceeds the user's preset permission region, the permission verification fails and the query is terminated. If the regional restriction keyword is not identified in the data query request, the regional relationship is confirmed as "cannot be determined", and the user's preset permission region is used as the query range by default, and the permission verification passes. If there are multiple regions corresponding to the regional restriction keywords, the region with the smaller determined region is injected as a variable into the subsequent process, and the permission verification passes.

[0057] In other embodiments, when the data query request is converted into target description text using a preset first prompt word template, structured standard keywords in the data query request are obtained. The standard keywords are used to indicate the name of a benchmark customer similar to the potential customer (such as "a company similar to Huawei"), the industry description of the potential customer (such as "a high-energy-consuming enterprise in Shandong Province"), or the name of the potential customer (such as "Industrial and Commercial Bank of China") in the data query request.

[0058] Fill the blanks in the first prompt word template with standard keywords to obtain the first prompt word; the first prompt word is used to indicate how to convert the standard keywords into standard text descriptions of the potential customer's industry and core business.

[0059] Input the first prompt word into the preset large language model to obtain the target description text output by the large language model.

[0060] In some embodiments, for benchmark customer similarity queries (such as "querying companies similar to Huawei"): by integrating the enterprise's internal customer information database with online search capabilities, a complete profile of the benchmark customer is constructed, and the large model is instructed to convert it into an accurate and standardized description of the customer's industry and core business.

[0061] In some embodiments, for specific queries on abstract concepts in a potential customer's industry (e.g., "querying high energy-consuming enterprises in Shandong Province"), a strategy combining domain knowledge enhancement and online retrieval is employed. First, the abstract concept is identified. Then, authoritative definitions and scopes are retrieved from government and enterprise domain knowledge bases and the internet. Finally, a large language model integrates information from multiple sources to output a corresponding, standardized list of industry enumerations.

[0062] In some embodiments, for semantically clear direct queries (such as "query ICBC"), no conversion is required and the query is directly allowed.

[0063] For example, "Qingdao Beer" can be translated as "a beer brewer and brand operator whose core business is the production and sale of beer"; "high energy-consuming enterprise" can be translated as "an enterprise engaged in the processing of petroleum, coal and other fuels, the manufacturing of chemical raw materials and chemical products, the non-metallic mineral products industry, the smelting and rolling of ferrous metals, the smelting and rolling of non-ferrous metals, and the production and supply of electricity, heat, gas and water".

[0064] In this embodiment, by performing intent-based security filtering and location-based permission verification on user-input data query requests, invalid and unauthorized queries by business personnel are avoided, thereby effectively saving system resources. By converting data query requests into target description text, it is ensured that the instructions flowing into the downstream query engine are standardized descriptions with clear structure and explicit semantics, laying a solid foundation for high-precision retrieval.

[0065] Figure 4 This is a flowchart illustrating another customer query method provided in an embodiment of this application. Figure 5 This is a schematic diagram of a prompt word template provided in an embodiment of this application. Figure 6 This is a schematic diagram of another prompt word template provided in an embodiment of this application. Figure 7 This is a schematic diagram illustrating yet another prompt word template provided in an embodiment of this application. (In conjunction with...) Figures 4 to 7As shown, in step S202 above, determining at least one target keyword based on the target description text and a preset prompt word template can be achieved through the following steps: S2021. Identify the customer type in the target description text.

[0066] The customer type is used to indicate whether a potential customer is a public sector entity such as a "government / public institution" or a private sector entity such as a "company".

[0067] S2022. In response to the customer type being a unit in the public sector, the preset second prompt word template is invoked to obtain target keywords about the potential customer's name.

[0068] In some embodiments, when the customer type indicates that the potential customer is a unit in the public sector, target keywords about the potential customer's name are obtained through a fuzzy keyword search method.

[0069] For example, such as Figure 5 As shown, when obtaining target keywords about potential customer names through fuzzy keyword retrieval, the target description text is filled into the fields to be filled in the second prompt word template to generate the second prompt word; the second prompt word is used to indicate the generation of the complete name of the customer that matches the description text.

[0070] In some embodiments, the second prompt word is input into a preset large language model to obtain at least one keyword related to the customer's name output by the large language model, and the keyword with the highest matching degree related to the customer's name is taken as the corresponding target keyword.

[0071] S2023. In response to a customer type of private sector entity, invoke at least one preset prompt word template to obtain at least one target keyword about the potential customer.

[0072] In some embodiments, when the customer type indicates that the potential customer is a unit in the private sector, a comprehensive search of target keywords about the multi-dimensional characteristics of the enterprise is obtained by combining customer name, national economic sector and business scope.

[0073] In some embodiments, the target description text is filled into the items to be filled corresponding to the second prompt word template, the preset third prompt word template, and the fourth prompt word template, respectively, to generate the corresponding second prompt word, third prompt word, and fourth prompt word; Among them, such as Figure 6As shown, the third suggestion word template is used to determine the industry of potential customers. The task of the third suggestion word template is to match the target description text with a pre-defined, known internal keyword library. Based on common sense, it outputs n target keywords that match the meaning of the target keywords, sorted in descending order of similarity, and finally provides a reason for matching the target keywords.

[0074] Due to the large number of field enumeration values ​​involved in the national economic sector, traditional small natural language processing models struggle to achieve accurate multi-word fuzzy semantic matching. This application addresses this pain point by employing a batch matching mechanism driven by a large language model. All enumeration value lists are submitted to the large language model in batches for judgment, thus circumventing the context length limitations of the large language model. Specifically, through a third-party prompt word template, the large language model is strictly instructed to output only semantically relevant standard enumeration values ​​from a given batch, and negative rule bases are inserted to prohibit creation or synonym substitution. Finally, the matching results from each batch are aggregated to form the final keyword set. This achieves accurate understanding of fuzzy industry concepts such as "high-tech" and "high-end manufacturing," and further improves overall reasoning efficiency through the batch strategy, achieving a balance between accuracy and efficiency.

[0075] Among them, such as Figure 7 As shown, the fourth prompt word template is used to extract and determine the business scope of potential customers from the domain knowledge database. Extracting the customer's business scope requires the large language model to reason based on common sense, generating professional and highly targeted target keywords, and explicitly requiring an explanation of the reasons for the selection to enhance interpretability. During the extraction process, a dynamically configurable "negative rule base" (e.g., specifying that "hotel industry" should not use "catering services") is invoked to filter the initial results, effectively avoiding common ambiguities within the domain and ensuring the accuracy of the generated target keywords.

[0076] In some embodiments, the second prompt word is used to instruct the generation of the full name of the customer that matches the target description text, the third prompt word is used to instruct the generation of the industry that matches the target description text, and the fourth prompt word is used to instruct the generation of the business scope that matches the target description text.

[0077] In some embodiments, the second prompt word, the third prompt word, and the fourth prompt word are respectively input into a preset large language model to obtain at least one keyword about the customer name, at least one keyword about the industry, and at least one keyword about the business scope output by the large language model.

[0078] Among at least one keyword related to the customer's name, the keyword with the highest matching degree is determined as the target keyword related to the potential customer's name; among at least one keyword related to the industry, the keyword with the highest matching degree is determined as the target keyword related to the potential customer's industry; among at least one keyword related to the business scope, the keyword with the highest matching degree is determined as the target keyword related to the potential customer's business scope.

[0079] In this embodiment of the application, different search modes are used for different types of customers to obtain target keywords corresponding to each type of customer, so as to fully capture the multi-dimensional characteristics of different types of customers and ensure the business relevance and comprehensiveness of the query results.

[0080] Figure 8 This is a flowchart illustrating yet another customer query method provided in an embodiment of this application. (Combined with...) Figure 8 As shown, in step S204 above, the initial customer list is screened for quality to obtain the target customer list needed by the user. This can be achieved through the following steps: S2041. Based on the target description text, obtain the first and second requirement descriptions of unstructured semantics.

[0081] In some embodiments, the first requirement describes conventional filtering conditions set by the user in the target description text, such as "filter out customers of technology companies," and the second requirement describes unconventional filtering conditions set by the user in the target description text, such as "exclude customers with operational risks in 2025."

[0082] S2042. Based on the first and second requirements descriptions, the initial customer list is subjected to the quality screening described above to obtain the target customer list.

[0083] In some embodiments, the initial customer list is preliminarily screened according to the first requirement description to obtain a preliminary customer list; the preliminary screening is used to indicate the filtering of customer information that does not meet the regular screening criteria.

[0084] For example, a pre-defined customer filtering prompt template command language model acts as a "data filtering assistant," not only performing keyword matching but also combining common sense and business logic to determine whether each piece of customer information in the initial customer list meets the first requirement description. For instance, it can accurately identify companies engaged in software development, integrated circuits, etc., from the description of "technology-based companies," rather than just customers whose names contain the word "technology." Finally, the filtering results, including customer identity information, are returned in standard JSON (JavaScript Object Notation) format, and a temporary result table is generated based on the filtering results as the initial customer list.

[0085] In some embodiments, a search engine query instruction is generated based on the second requirement description (such as "whether it is a specialized and innovative enterprise" or "whether it has recently experienced a financial crisis"). The search engine query instruction is then used to retrieve publicly available information (such as news, announcements, financial reports, etc.) of each customer in the initial screening customer list. In other words, external information is used to verify and enhance the customer information in the initial screening customer list.

[0086] In some embodiments, the initial customer list is validated and filtered based on the publicly available information of each customer to obtain a target customer list; the validation and filtering is used to indicate the filtering of customer information that does not meet the non-standard filtering conditions.

[0087] In other embodiments, the filtered customer information, i.e. the reason for filtering, is recorded to enhance the interpretability of the resulting target customer list.

[0088] In this embodiment, the initial customer list retrieved is subjected to secondary screening and data augmentation. Traditional SQL queries are insufficient to handle complex and ambiguous semantic requirements. This embodiment combines a collaborative verification mechanism with screening and verification logic to achieve quality screening and verification of customer information in the initial customer list, thereby further improving the accuracy of potential customer queries.

[0089] As can be seen, the above mainly describes the solutions provided by the embodiments of this application from a methodological perspective. To achieve the above functions, the embodiments of this application provide corresponding hardware structures and / or software modules for executing each function. Those skilled in the art should readily recognize that, in conjunction with the modules and algorithm steps of the various examples described in the embodiments disclosed herein, the embodiments of this application can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed by hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this invention.

[0090] This application embodiment can divide the customer query device into functional modules according to the above method example. For example, each function can be divided into its own functional module, or two or more functions can be integrated into one processing module. The integrated module can be implemented in hardware or as a software functional module. Optionally, the module division in this application embodiment is illustrative and only represents one logical functional division; other division methods may be used in actual implementation.

[0091] Figure 9 This is a schematic diagram of the structure of an electronic device provided in an embodiment of this application. Figure 9As shown, the electronic device includes: a processor 902, a communication interface 903, and a bus 904. Optionally, the electronic device may also include a memory 901.

[0092] Processor 902 may implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 902 may be a central processing unit, a general-purpose processor, a digital signal processor, an application-specific integrated circuit (ASIC), a field-programmable gate array (FPGA), or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It may implement or execute various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this application. Processor 902 may also be a combination of functions implementing computing capabilities, such as a combination of CPU0 and CPU1, a DSP, and a microprocessor.

[0093] The communication interface 903 includes a receiving unit and a transmitting unit, and is used to connect with other devices via a communication network. This communication network can be Ethernet, a wireless access network, a wireless local area network (WLAN), etc.

[0094] The memory 901 may be a read-only memory (ROM) or other type of static storage device capable of storing static information and instructions, random access memory (RAM) or other type of dynamic storage device capable of storing information and instructions, or electrically erasable programmable read-only memory (EEPROM), disk storage media or other magnetic storage devices, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and accessible by a computer, but is not limited thereto.

[0095] In one possible implementation, the memory 901 can exist independently of the processor 902. The memory 901 can be connected to the processor 902 via a bus 904 and is used to store instructions or program code. When the processor 902 calls and executes the instructions or program code stored in the memory 901, it can implement the client query method provided in this embodiment of the invention.

[0096] In another possible implementation, the memory 901 can also be integrated with the processor 902.

[0097] The 904 bus can be an extended industry standard architecture (EISA) bus, etc. The 904 bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 9 The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0098] Through the above description of the implementation methods, those skilled in the art can clearly understand that, for the sake of convenience and brevity, only the division of the above functional modules is used as an example. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the service calling device can be divided into different functional modules to complete all or part of the functions described above.

[0099] This application also provides a computer-readable storage medium. All or part of the processes in the above method embodiments can be executed by computer instructions instructing related hardware. The program can be stored in the computer-readable storage medium, and when executed, it can include the processes of the above method embodiments. The computer-readable storage medium can be any of the foregoing embodiments or memory. The computer-readable storage medium can also be an external storage device of the electronic device, such as a plug-in hard drive, smart media card (SMC), secure digital (SD) card, flash card, etc., equipped on the electronic device. Further, the computer-readable storage medium can include both internal storage units and external storage devices of the electronic device. The computer-readable storage medium is used to store the computer program and other programs and data required by the electronic device. The computer-readable storage medium can also be used to temporarily store data that has been output or will be output.

[0100] This application also provides a computer program product, which includes a computer program that, when run on a computer, causes the computer to execute any of the customer query methods provided in the above embodiments.

[0101] The above are merely specific embodiments of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions within the technical scope disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A customer query method, characterized in that, include: Retrieve the target description text to be queried; The target description text is used to indicate a standardized description of a potential customer; Based on the target description text and the preset prompt word template, at least one target keyword is determined; the target keyword is used to indicate semantic information that matches the characteristics of the potential customer; Based on the at least one target keyword, an initial customer list is obtained; the initial customer list is used to indicate customer information obtained from a local customer database. The initial customer list is subjected to quality screening to obtain the target customer list required by the user; the quality screening is used to instruct the filtering of customer information in the initial customer list in combination with the unstructured semantics in the target description text.

2. The method according to claim 1, characterized in that, The process of obtaining the target description text to be queried includes: Obtain the data query request input by the user; the data query request is used to indicate a vague description about the potential customer; Upon successful permission verification of the data query request, the data query request is converted into the target description text using a preset first prompt word template. The permission verification is used to indicate whether the geographical restriction keywords in the data query request meet the geographical range that the user can query, and the first prompt word template is used to transform the data query request from a vague description into a standardized description.

3. The method according to claim 2, characterized in that, The step of converting the data query request into the target description text using a preset first prompt word template includes: Obtain structured standard keywords from the data query request. The standard keywords are used to indicate the name of a benchmark customer similar to the potential customer, the industry description of the potential customer, or the name of the potential customer in the data query request. The standard keywords are filled into the fields to be filled in the first prompt word template to obtain the first prompt word; the first prompt word is used to indicate that the standard keywords are converted into a standard text description of the potential customer's industry and core business. The first prompt word is input into a preset large language model to obtain the target description text output by the large language model.

4. The method according to claim 1, characterized in that, The step of determining at least one target keyword based on the target description text and a preset prompt word template includes: Identify the customer type of the target description text, the customer type being used to indicate whether the potential customer is an entity in the public sector or an entity in the private sector; In response to the customer type being a unit in the public sector, a preset second prompt word template is invoked to obtain the target keywords related to the potential customer's name. The second prompt word template is used to determine the full name of the potential customer. In response to the customer type being an entity in the private sector, at least one preset prompt word template is invoked to obtain at least one target keyword about the potential customer. The at least one target keyword includes target keywords about the potential customer's name, target keywords about the potential customer's industry, and target keywords about the potential customer's business scope.

5. The method according to claim 4, characterized in that, The step of calling a preset second prompt word template to obtain at least one target keyword related to the customer's name includes: The target description text is populated into the fields to be filled in the second prompt word template to generate a second prompt word; the second prompt word is used to indicate the generation of the full name of the customer that matches the description of the target description text. The second prompt word is input into a preset large language model to obtain at least one keyword related to the customer's name output by the large language model, and the keyword with the highest matching degree related to the customer's name is taken as the corresponding target keyword.

6. The method according to claim 4, characterized in that, The process of retrieving at least one target keyword about the potential customer by calling at least one preset prompt word template includes: The target description text is filled into the corresponding fields to be filled in the second prompt word template, the preset third prompt word template, and the fourth prompt word template, respectively, to generate the corresponding second prompt word, third prompt word, and fourth prompt word; The third prompt word template is used to determine the industry of the potential customer, and the fourth prompt word template is used to determine the business scope of the potential customer. The second prompt word is used to instruct the generation of a complete name that matches the customer described in the target description text; the third prompt word is used to instruct the generation of an industry that matches the customer described in the target description text; and the fourth prompt word is used to instruct the generation of a business scope that matches the customer described in the target description text. The second, third, and fourth prompt words are input into a preset large language model to obtain at least one keyword related to the customer name, at least one keyword related to the industry, and at least one keyword related to the business scope output by the large language model. Among the at least one keyword related to the customer's name, the keyword with the highest matching degree is determined as the target keyword related to the potential customer's name; among the at least one keyword related to the industry, the keyword with the highest matching degree is determined as the target keyword related to the potential customer's industry; among the at least one keyword related to the business scope, the keyword with the highest matching degree is determined as the target keyword related to the potential customer's business scope.

7. The method according to claim 1, characterized in that, The process of obtaining an initial customer list based on the at least one target keyword includes: Based on the at least one target keyword, a target query statement is generated; the target query statement is a structured query statement. The target query statement is used to query matching customer information in the local customer database; wherein, the local customer database pre-stores multiple customer information, including customer name, customer industry and customer business scope; Based on the matched customer information, the initial customer list is generated.

8. The method according to claim 1, characterized in that, The process of performing quality screening on the initial customer list to obtain the target customer list that the user needs to query includes: Based on the target description text, obtain a first requirement description and a second requirement description of unstructured semantics; the first requirement description is used to indicate the regular filtering conditions set by the user, and the second requirement description is used to indicate the non-regular filtering conditions set by the user. The initial customer list is subjected to quality screening based on the first and second requirements descriptions to obtain the target customer list.

9. The method according to claim 8, characterized in that, The process of performing quality screening on the initial customer list based on the first and second requirement descriptions to obtain the target customer list includes: Based on the first requirement description, the initial customer list is initially screened to obtain a preliminary customer list; the preliminary screening is used to indicate the filtering of customer information that does not meet the regular screening conditions. Based on the second requirement, a search engine query instruction is generated, and the public information of each customer in the initial screening customer list is retrieved online through the search engine query instruction. Based on the publicly available information of each customer, the initial screening customer list is verified and filtered to obtain the target customer list; the verification and filtering is used to indicate the filtering of customer information that does not meet the non-standard screening conditions.

10. An electronic device, characterized in that, It includes a processor and a memory, the processor being coupled to the memory; the memory is used to store computer instructions, which are loaded and executed by the processor to enable the computer device to implement the customer query method as described in any one of claims 1 to 9.

11. A computer-readable storage medium, characterized in that, The computer-readable storage medium includes computer-executable instructions that, when executed on a computer, cause the computer to perform the customer query method as described in any one of claims 1 to 9.

12. A computer program product, characterized in that, The computer program product includes a computer program that, when run on an electronic device, causes the electronic device to perform the customer query method as described in any one of claims 1 to 9.

Images