Product marketing method and device for realizing Agent based on RPA, AI and LLM, equipment and storage medium

Abstract

The invention provides a product marketing method, device and equipment for realizing an agent based on robot process automation RPA, artificial intelligence AI and a large language model LLM, and a storage medium. The product marketing method comprises the steps that key features related to commercial cooperation are extracted from trade related data of potential customers, the cooperation value of the potential customers is determined based on the key features, and the key features comprise the number of product orders, the amount of transaction and / or the number of imports and exits; the potential customers with the cooperation values meeting the requirements serve as to-be-mined customers, and product promotion content matched with the to-be-mined customers is generated; sending the product promotion content to the corresponding to-be-mined customer; feedback content of the customer to be mined about the product promotion content is identified to determine a cooperation intention. Through adoption of the technical scheme, the problem of low efficiency of manual follow-up of the potential customers is solved, and the follow-up efficiency and the cooperation success rate of the potential customers are effectively improved.

Description

Technical Field

[0001] This application relates to the field of customer mining, and in particular to a product marketing method, apparatus, equipment and storage medium based on RPA, AI and LLM to realize agents. Background Technology

[0002] Robotic Process Automation (RPA) uses specific "robot software" to simulate human operations on a computer and automatically execute process tasks according to rules.

[0003] Artificial intelligence (AI) is a technical science that studies and develops theories, methods, technologies, and application systems for simulating, extending, and expanding human intelligence.

[0004] Large Language Models (LLMs) are models trained on massive amounts of text that can recognize human language, perform language-related tasks, and have a large number of parameters.

[0005] Artificial Intelligence Agents (AI Agents) are capable of perceiving their environment, making decisions, and executing actions. Unlike traditional artificial intelligence, they possess the ability to think and act independently, and can utilize tools to achieve given goals. AI Agents are based on Large Language Models (LLMs) as their core computing engine, enabling them to engage in dialogue, perform tasks, reason, and exhibit a degree of autonomy. They possess the ability to autonomously understand, perceive, plan, remember, and use tools, and can automate complex tasks. Specifically, LLM-driven AI Agents, composed of various AI capabilities, can interact with employees using natural language, understand employee instructions and needs, and provide feedback and responses; they can acquire domain-specific knowledge relevant to the business to complete complex professional tasks; they can break down complex tasks into several executable tasks and use data and tools to complete them; they can also collaborate with employees, and AI Agents can collaborate with each other to complete complex tasks, enabling digital employees to leap from automation to intelligence, helping employees complete their work more efficiently, and fully realizing human-machine collaboration.

[0006] In related technologies, customer acquisition and maintenance processes all require manual operation by sales personnel. During the process of identifying potential customers, sales personnel not only need to systematically and comprehensively introduce the products or services to be sold based on the potential customers' interests, but also need to follow up frequently, such as conducting regular telephone follow-ups or continuously sending product brochures, success stories, and promotional information. Furthermore, sales personnel need to identify high-value customers worthy of in-depth exploration based on customer feedback and provide further product introductions and follow-ups to these customers. The process of identifying each potential customer is not only tedious but also time-consuming. When the number of potential customers is large, manually identifying and maintaining a large number of customers through sales personnel is extremely time-consuming, labor-intensive, and inefficient. Summary of the Invention

[0007] This application provides a product marketing method, apparatus, device, and storage medium based on RPA, AI, and LLM to address the problems existing in related technologies. The technical solution is as follows: In a first aspect, embodiments of this application provide a product marketing method based on RPA, AI, and LLM to implement an agent, including: Key features related to business cooperation are extracted from the trade-related data of each potential customer, and the cooperation value of each potential customer is determined based on the key features. The trade-related data includes company introduction, contact information and / or product import and export information, and the key features include product order quantity, transaction amount and / or import and export quantity. Potential customers whose cooperation value meets the requirements are identified as potential customers to be explored, and product promotion content is generated to match each potential customer. Send product promotional content to the relevant potential customers; Identify feedback from potential clients regarding the promotional content for the corresponding products to determine their willingness to cooperate.

[0008] In one embodiment, the method provided by the present invention further includes: Obtain basic information about each potential customer from the customer management platform. This basic information includes the company name and country of origin. Based on basic information, trade-related data of potential customers are retrieved from multiple information query platforms and entered into the customer management platform.

[0009] In one implementation, potential customers include: Historical clients with initial cooperation intentions found on the customer management platform, and... Customers who have cooperative relationships with various competitors and / or customers who have transactions with competitors, as determined by competitor data and / or competitor product data. Accordingly, the method provided by the present invention further includes: Customer basic information on customers with whom we have cooperative relationships and customers with whom we have transactions with competing products will be entered into the customer management platform.

[0010] In one implementation, trade-related data of potential customers is retrieved from multiple information query platforms based on basic information, including: The system invokes multiple agents that have been set up to retrieve various types of trade-related data related to the basic information of each potential customer from different information query platforms. Each agent includes: The first agent is used to log in to the first information retrieval platform to search for company profiles related to the basic information of potential customers. The second agent is used to log in to the second information retrieval platform to search for contact information related to the basic information of potential customers. The third agent is used to log in to the third information retrieval platform to search for product import and export information related to the basic information of potential customers.

[0011] In one implementation, trade-related data of potential customers is retrieved from multiple information query platforms based on basic information, including: By logging into different information query platforms through RPA, various types of trade-related data related to the basic information of each potential customer can be found.

[0012] In one implementation, key features relevant to business cooperation are extracted from the trade-related data of each potential customer, and the cooperation value of each potential customer is determined based on these key features, including: The established fourth agent is invoked to analyze the trade-related data of each potential customer, thereby obtaining key features and their feature values ​​in different dimensions related to business cooperation. The fourth agent performs a weighted summation of multiple key feature values ​​to obtain a comprehensive score for the potential customer, which is used to evaluate the value of cooperation.

[0013] In one implementation, generating product marketing content tailored to each target customer includes: The established fifth agent is invoked to identify the validity of each customer email address to be mined. For each potential customer with a valid email address, the fifth agent analyzes the customer's trade-related data to determine the types of products the customer is interested in and obtains product images. The fifth agent selects the target template corresponding to the product type from multiple pre-configured email templates, and adds product-related data of interest to the customer to be mined into the target template to obtain product promotion content. Accordingly, product promotional materials will be sent to the relevant potential customers, including: The fifth agent edits the product promotion content into the language commonly used by the target customers and sends it to them via email. The language commonly used by the target customers is determined through trade-related data.

[0014] In one implementation, during the construction of the fifth agent, various types of personalized adjustment parameters are set for the fifth agent, and the parameter values ​​of each personalized adjustment parameter are configured to maximize the personalization of the output results of the fifth agent. The personalized adjustment parameters include a first parameter, a second parameter, and a third parameter. The first parameter is used to adjust the level of detail of the output information, the second parameter is used to adjust the randomness of the output information, and the third parameter is used to adjust the selection range of product-related data in the product promotion content.

[0015] In one implementation, identifying feedback from potential clients regarding product promotion content to determine their cooperation intentions includes: Upon receiving a reply email from a customer to be identified, the established sixth agent is invoked to filter emails of the automatically replied type. For emails that are not automatically replied to, the sixth agent performs semantic recognition on the feedback content of each potential customer using natural language processing (NLP) to determine the degree of cooperation intention of each potential customer. The sixth agent is trained during the configuration process using the following prompts: The rules for judging automatic reply emails, the criteria for classifying the degree of intent, and historical email interaction cases between the current sales party and its partners.

[0016] In one implementation, identifying feedback from potential clients regarding product promotion content to determine their cooperation intentions includes: Periodically identify the receipt status of preset sending email addresses to track feedback from potential customers regarding product promotion content; Upon receiving feedback from any potential client, the system invokes Natural Language Processing (NLP) services to perform semantic recognition on the feedback content, thereby determining the client's cooperation intentions.

[0017] Secondly, embodiments of this application provide a product marketing device based on RPA, AI, and LLM to implement an agent, the device comprising: The cooperation value determination module is configured to extract key features related to business cooperation from the trade-related data of each potential customer, and determine the cooperation value of each potential customer based on the key features. The trade-related data includes company introduction, contact information and / or product import and export information, and the key features include product order quantity, transaction amount and / or import and export quantity. The product promotion content generation module is configured to identify potential customers whose cooperation value meets the requirements as potential customers to be explored, and generate product promotion content that matches each potential customer. The product promotion content sending module is configured to send product promotion content to the corresponding potential customers; The cooperation intention determination module is configured to identify feedback from potential clients regarding product promotion content in order to determine their cooperation intention.

[0018] In one embodiment, the apparatus provided by the present invention further includes: The basic information acquisition module is configured to acquire basic information about potential customers from the customer management platform. This basic information includes the company name and country of origin. The trade-related data query module is configured to search for trade-related data of potential customers from multiple information query platforms based on basic information and enter it into the customer management platform.

[0019] In one implementation, potential customers include: Historical clients with initial cooperation intentions found on the customer management platform, and... Customers who have cooperative relationships with various competitors and / or customers who have transactions with competitors, as determined by competitor data and / or competitor product data. Accordingly, the apparatus provided in this embodiment of the invention further includes: Basic information on customers with whom you have cooperative relationships and customers with whom you have transactions with competing products will be entered into the customer management platform.

[0020] In one implementation, the trade-related data query module includes: The Agent invocation unit invokes multiple agents that have been set up to retrieve various types of trade-related data related to the basic information of each potential customer from different information query platforms. Each agent includes: The first agent is used to log in to the first information retrieval platform to search for company profiles related to the basic information of each potential customer. The second agent is used to log in to the second information retrieval platform to search for contact information related to the basic information of each potential customer. The third agent is used to log in to the third information retrieval platform to search for product import and export information related to the basic information of each potential customer.

[0021] In one implementation, the trade-related data query module includes: The RPA task execution unit is used to log in to different information query platforms via RPA to find various types of trade-related data related to the basic information of each potential customer.

[0022] In one implementation, the cooperation value determination module is specifically configured as follows: The established fourth agent is invoked to analyze the trade-related data of each potential customer, thereby obtaining key features and their feature values ​​in different dimensions related to business cooperation. The fourth agent performs a weighted summation of multiple key feature values ​​to obtain a comprehensive score for the potential customer, which is used to evaluate the value of cooperation.

[0023] In one implementation, the product promotion content generation module is specifically configured as follows: The established fifth agent is invoked to identify the validity of each customer email address to be mined. For each potential customer with a valid email address, the fifth agent analyzes the customer's trade-related data to determine the types of products the customer is interested in and obtains product images. The fifth agent selects the target template corresponding to the product type from multiple pre-configured email templates, and adds product-related data of interest to the customer to be mined into the target template to obtain product promotion content. Accordingly, the product promotion content sending module is specifically configured as follows: The fifth agent edits the product promotion content into the language commonly used by the target customers and sends it to them via email. The language commonly used by the target customers is determined through trade-related data.

[0024] In one implementation, during the construction of the fifth agent, various types of personalized adjustment parameters are set for the fifth agent, and the parameter values ​​of each personalized adjustment parameter are configured to maximize the personalization of the output results of the fifth agent. The personalized adjustment parameters include a first parameter, a second parameter, and a third parameter. The first parameter is used to adjust the level of detail of the output information, the second parameter is used to adjust the randomness of the output information, and the third parameter is used to adjust the selection range of product-related data in the product promotion content.

[0025] In one implementation, the cooperation intention determination module is configured to: Upon receiving a reply email from a customer to be identified, the established sixth agent is invoked to filter emails of the automatically replied type. For emails that are not automatically replied to, the sixth agent performs semantic recognition on the feedback content of each potential customer using natural language processing (NLP) to determine the degree of cooperation intention of each potential customer. The sixth agent is trained during the configuration process using the following prompts: The rules for judging automatic reply emails, the criteria for classifying the degree of intent, and historical email interaction cases between the current sales party and its partners.

[0026] In one implementation, the cooperation intention determination module is configured to: Periodically identify the receipt status of preset sending email addresses to track feedback from potential customers regarding product promotion content; Upon receiving feedback from any potential client, the system invokes Natural Language Processing (NLP) services to perform semantic recognition on the feedback content, thereby determining the client's cooperation intentions.

[0027] Thirdly, embodiments of this application provide an electronic device comprising a memory and a processor. The memory and the processor communicate with each other via an internal connection path. The memory stores instructions, and the processor executes the instructions stored in the memory. When the processor executes the instructions stored in the memory, it causes the processor to perform the method described in any of the above embodiments.

[0028] Fourthly, embodiments of this application provide a computer-readable storage medium that stores a computer program, wherein when the computer program is run on a computer, the methods in any of the above-described embodiments are executed.

[0029] The advantages or beneficial effects of the above technical solutions include at least the following: By using intelligent agents to analyze trade-related data of potential customers to determine their cooperation value, this not only solves the problem of low efficiency in manual data analysis by sales personnel, but also effectively identifies high-value customers, ensuring customer quality and avoiding delays in follow-up due to inaccurate identification of high-value customers. By using intelligent agents to replace manual labor in customizing product promotion content for potential customers, this not only reduces labor and time costs but also achieves efficient and precise promotion of products. Furthermore, by using intelligent agents to replace manual labor in identifying feedback from potential customers and determining the degree of cooperation intention of each customer, this not only solves the problem of time and energy constraints in manually reviewing customer responses one by one, but also ensures the accuracy of customer cooperation intention assessments. Especially when the number of potential customers is large, this significantly improves the processing efficiency of customer feedback, avoids the loss of high-intent customers due to delays in manual processing, and increases the success rate of cooperation with potential customers.

[0030] The above overview is for illustrative purposes only and is not intended to be limiting in any way. In addition to the illustrative aspects, embodiments, and features described above, further aspects, embodiments, and features of this application will become readily apparent from the accompanying drawings and the following detailed description. Attached Figure Description

[0031] In the accompanying drawings, unless otherwise specified, the same reference numerals throughout the various drawings denote the same or similar parts or elements. These drawings are not necessarily drawn to scale. It should be understood that these drawings depict only some embodiments disclosed in this application and should not be construed as limiting the scope of this application.

[0032] Figure 1a This is a flowchart of a product marketing method based on RPA, AI, and LLM to implement an agent, as provided in Embodiment 1 of this application; Figure 1b A schematic diagram of the AI ​​search component provided in Embodiment 1 of this application; Figure 1c This is a schematic diagram of the product promotion status provided in Embodiment 1 of this application; Figure 1d A schematic diagram of the AI ​​activity interface in a customer management platform is provided for Embodiment 1 of the present invention; Figure 2 A flowchart illustrating a product marketing method based on RPA, AI, and LLM for implementing an agent, provided in Embodiment 2 of this application; Figure 3 This is a structural block diagram of a product marketing device based on RPA, AI, and LLM to implement an agent, as provided in Embodiment 3 of this application. Figure 4 This is a structural block diagram of an electronic device provided in Embodiment 4 of this application. Detailed Implementation

[0033] The embodiments of this application are described in detail below. Examples of these embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain this application, and should not be construed as limiting this application.

[0034] In the description of this application, the term "multiple" means two or more.

[0035] In the description of this application, the terms "first," "second," "third," "fourth," "fifth," and "sixth" are used only to distinguish different agents and do not have any limiting effect.

[0036] In the description of this application, "trade-related data" includes domestic trade-related data and import / export trade-related data. This "related data" includes company information, contact information, and / or product import / export information. The company information includes the company name, website, address, contact information, product category, and product description. Contact information includes the contact person's name, telephone number, and / or email address. Product import / export information includes descriptions of the imported / exported products, importer's name, exporter's name, number of imports / exports, quantity, weight, value, transaction method, mode of transport, and country of origin.

[0037] In the description of this application, "key features related to business cooperation" refers to data related to product transactions during the business cooperation process, such as product order quantity, transaction amount and / or import and export quantity, etc.

[0038] In the description of this application, "potential customer" refers to an undeveloped customer who has a demand for and purchasing power for a certain type of product (or service) offered by the seller. These customers present sales opportunities with the company, and through the efforts of the company and its sales personnel, they can be converted into cooperative customers. "Potential customers" represent a broad customer group in the initial stage of the conversion process.

[0039] In the description of this application, "potential customers" refers to customers with cooperative value selected from potential customers.

[0040] In the description of this application, the "customer management platform" is a centralized, digital customer information database and sales process facilitator. All customer information is stored in the "customer management platform," including information on customers with whom cooperation has already been established, customers with a clear intention to cooperate, and potential customers.

[0041] In the description of this application, "competitor data" refers to product data, sales data, etc. of all peer companies in the same industry that have business relationships with it (may cooperate, complement, or have no direct conflict), or other companies that provide similar products or services.

[0042] In the description of this application, "competitive product data" refers to product or service data that competes with the current seller's products in the market, and can be substituted for each other or meet the same user needs, including service targets, target customers, etc.

[0043] In the description of this application, "first information retrieval platform" refers to a website used to retrieve information related to company introductions.

[0044] In the description of this application, "second information retrieval platform" refers to a website used to search for contact information.

[0045] In the description of this application, the "third information retrieval platform" is a website used to search for product import and export data.

[0046] In this application, the term "Natural Language Processing" (NLP) is an important research direction in the field of artificial intelligence. It integrates knowledge from multiple disciplines such as linguistics, computer science, machine learning, mathematics, and cognitive psychology. It is an interdisciplinary field combining computer science, artificial intelligence, and linguistics, encompassing two main aspects: natural language understanding and natural language generation. Its research content includes multiple levels such as characters, words, phrases, sentences, paragraphs, and texts, serving as a bridge between machine language and human language. In this application, NLP services are invoked to perform semantic understanding of the feedback content from the client to be mined, in order to determine the client's cooperation intent.

[0047] In the description of this application, the term "cue word" is a natural language instruction used to guide an agent to perform a specific task, transforming general AI into a powerful tool for solving specific problems by clearly defining the agent's role, capabilities, and behavioral boundaries.

[0048] In the description of this application, the term "customer lead pool" refers to a centralized repository of contact information and data of all potential customers that an enterprise has collected through various channels but has not yet converted into paying customers. It is a dynamic resource pool that requires continuous nurturing and management.

[0049] In the description of this application, the term "product knowledge base" is a database that centrally stores, manages, and shares information related to the entire product lifecycle. It can provide a unified information source for various teams within an enterprise and support external customer service. It is a core tool for enterprises to improve collaboration efficiency and ensure business standardization.

[0050] In the description of this application, the term "field" refers to the name of a functional module or functional unit in the customer management platform, such as "company name", "email", "telephone", etc., and each type of field has a corresponding field value.

[0051] In the description of this application, the term "emotional tendency keywords" refers to words or expressions in a customer's response that reflect their subjective emotions, attitudes, or sentiments, and are primarily used to determine the customer's level of enthusiasm, resistance, or neutrality toward cooperation.

[0052] In the description of this application, the term "backend service" itself refers to a program running on a server (such as a digital employee platform), which acts as the "central brain" and "dispatch center" of the intelligent agent system. In the specific scenario of product marketing, this program continuously analyzes user data and, based on its internal logic, intelligently schedules and connects multiple intelligent agents (i.e., dedicated program modules) to work sequentially, such as customer data completion, high-value customer analysis, product promotion email sending, and customer intent identification. Ultimately, it accurately identifies the customer's level of intent and provides the identification results to business personnel.

[0053] In the description of this application, the implementation of the Digital Employee Platform (WEP) has gone through three stages. The first stage is the automation stage: for RPA stages with very low business complexity, software automation technology is used to automate rule-based, predefined procedural tasks. The second stage is the intelligence stage: extending the boundaries of RPA with the help of AI, such as processing unstructured documents and making judgments and decisions based on data. The third stage is the human-machine collaboration stage: utilizing the understanding, planning, and execution capabilities of large models to automate complex tasks end-to-end.

[0054] In the human-machine collaboration phase, the digital employee platform serves as a bridge connecting workers and systems, workers and data, and systems and data. It is capable of: operating complex systems, processing various types of data, and interacting and collaborating with employees. The digital employee platform helps industries build large-scale, model-enabled digital employees (i.e., intelligent agents), achieving automation, intelligence, and human-machine collaboration in business processes.

[0055] The digital employee platform can seamlessly integrate multiple capabilities such as Agentic Process Automation (APA), Agentic Document Processing (ADP), and Agentic Business Insights (ABI). It has five major functions: "business understanding", "process creation", "run anywhere", "centralized management and control" and "human-machine collaboration". It enables enterprises to achieve end-to-end intelligent automation of business processes, replace manual operations, further improve business efficiency, and accelerate digital transformation.

[0056] These and other aspects of the embodiments of this application will become clear from the following description and accompanying drawings. In these descriptions and drawings, some specific implementations of the embodiments of this application are specifically disclosed to illustrate some ways of carrying out the principles of the embodiments of this application; however, it should be understood that the scope of the embodiments of this application is not limited thereto. Rather, the embodiments of this application include all variations, modifications, and equivalents falling within the spirit and scope of the appended claims.

[0057] The following description, in conjunction with the accompanying drawings, provides a detailed overview of the product marketing method, apparatus, device, and storage medium based on RPA, AI, and LLM for implementing agents, as provided in the embodiments of this application.

[0058] Example 1 First, it should be noted that the product marketing method based on RPA, AI, and LLM provided in this embodiment of the invention can be applied to customer mining and sales of various types of products, such as customer mining and product promotion of food, cosmetics, clothing, etc. This embodiment does not make specific limitations on this.

[0059] The product marketing method based on RPA, AI, and LLM provided in this embodiment of the invention can be implemented by building an agent. This agent can be a single agent or multiple agents. If multiple agents are used, each agent has a different function, and each agent is used to execute different steps of the product marketing method. This embodiment does not specifically limit the number of agents. The agents can be developed using an AI agent development platform and deployed on a digital employee platform. It is understood that during the agent building process, a large number of system-level prompts need to be configured for training, and some callable tools need to be configured for the agents. In the product marketing application scenario of this invention, the prompts include: natural language rules, the specific content of each step executed by the agent, such as extracting key features related to business cooperation from trade-related data, evaluation criteria for determining the cooperation value of each potential customer based on key features, judgment criteria for customers to be mined, email templates for generating product promotion content, customer intent level grading standards, product knowledge bases, etc. Callable tools include: interfaces corresponding to different information query platforms, interfaces of email address validity recognition platforms, product database interfaces, etc. When the product marketing method provided in this embodiment of the invention is executed through a single Agent, the aforementioned prompts and tools can be configured for the same Agent. After the Agent is trained, the product marketing method provided in this embodiment of the invention can be implemented by calling the Agent. When the product marketing method provided in this embodiment is executed through multiple Agents, during the training process, the aforementioned prompts and tools can be assigned to different Agents according to the functions to be implemented by each Agent. After each Agent is trained, steps S1 to S3 provided in this embodiment can be executed by calling different Agents respectively.

[0060] The technical solutions provided by the embodiments of the present invention will be described in detail below.

[0061] Figure 1a This is a flowchart of a product marketing method based on RPA, AI, and LLM to implement an agent, as provided in Embodiment 1 of this application. Figure 1a As shown, the method may include the following steps: S110: Extract key features relevant to business cooperation from the trade-related data of each potential customer, and determine the cooperation value of each potential customer based on the key features.

[0062] Understandably, potential customers can be comprised of two groups. One group consists of existing customers with initial cooperation intentions. These existing customers could be those who have previously inquired about a particular product or service from the current seller, or those with whom the salesperson has had initial communication. Basic information about these customers can be maintained on the customer management platform. The other group of potential customers can be those identified through competitor data and / or rival product data, who have existing partnerships with other competitors, and / or have transactions with competing products.

[0063] For example, the intelligent agent can obtain customers with cooperative relationships with each competitor based on the list of competitors provided by the seller, and can obtain customers with transactions involving competing products based on the list of competitors. The customers obtained by the intelligent agent can be used as candidate customers. By matching candidate customers with historical customers already stored on the customer management platform, it can be determined whether a candidate customer is included in the historical customer list. Specifically, a similarity matching method can be used to determine whether the candidate customer's basic information exists in the customer management platform. That is, if the similarity between a candidate customer's name and a historical customer's name reaches a set threshold, it is determined that the basic information of the candidate customer is stored in the customer management platform; if the similarity between a candidate customer and a historical customer does not reach the set threshold, it is determined that the candidate customer is a new customer different from the customers already stored on the customer management platform. In this case, the intelligent agent needs to enter the basic information of the candidate customer into the customer management platform.

[0064] In this embodiment, by expanding the number of potential customers maintained on the customer management platform, sales opportunities are directly increased, injecting resources for the company's sustainable development. By utilizing intelligent agents to expand the number of potential customers, the inefficiency of sales personnel manually searching for customers is avoided, completely freeing manpower from repetitive labor and improving the efficiency of potential customer growth.

[0065] In this embodiment, trade-related data includes domestic trade-related data and import / export trade-related data. This data includes company information, contact information, and / or product import / export information. The company information includes the company name, website, address, contact information, product category, and product description. Contact information includes the contact person's name, phone number and / or email address, and job title. Product import / export information includes product descriptions, importer's name, exporter's name, number of imports / exports, quantity, weight, value, transaction method, mode of transport, and country of origin.

[0066] For example, trade-related data for each potential customer can be obtained based on their basic information, including company name and country of origin. This trade-related data can be stored as supplementary information for potential customers in the customer lead pool of the customer management platform. The process for completing the trade-related data for each potential customer can be triggered as follows: the backend service can trigger the completion process whenever it detects new customer basic information being entered into the customer management platform; alternatively, it can be triggered periodically via RPA to complete or update the trade-related data for potential customers. This embodiment does not specifically limit this approach. The backend service can be a program written in a programming language (e.g., Python).

[0067] For example, supplementing trade-related data based on the basic information of each potential customer can be achieved in the following way: First, by invoking intelligent agents, basic information of potential customers can be retrieved from the customer management platform. Based on this information, the intelligent agents can then invoke pre-configured information query tools to find trade-related data for each potential customer. These information query tools include interfaces corresponding to various information query platforms.

[0068] Second, pre-build RPA processes and use RPA to log in to different information query platforms to find trade-related data of potential customers.

[0069] In this embodiment, by supplementing the information of potential customers, basic customer data is upgraded to complete, accurate, and comprehensive customer data. Supplementing customer data not only helps to identify high-value customers more quickly but also helps to clarify the core needs of different customers, making subsequent customer acquisition and follow-up more accurate and efficient, thereby accelerating the cooperation process with customers. By using intelligent agents or RPA to find trade-related data of potential customers and inputting it into the customer management platform, the information of a massive number of potential customers can be supplemented in a short time. Compared with the manual information retrieval and input methods in related technologies, the automatic supplementation method of potential customer information provided in this embodiment effectively improves the efficiency and accuracy of information retrieval.

[0070] Furthermore, to improve the targeting, accuracy, and speed of trade-related data searches, multiple agents can be pre-built for searching different types of trade-related data. When customer data needs to be supplemented, the pre-built agents are invoked, and each agent can search for various types of trade-related data related to potential customers from different information query platforms. Each agent has a different function; for example, the first agent searches for company introductions related to the basic information of potential customers on the first information retrieval platform; the second agent searches for contact information related to the basic information of potential customers on the second information retrieval platform; and the third agent searches for product import / export information related to the basic information of potential customers on the third information retrieval platform. In this embodiment, the retrieval process of different types of trade-related data by each agent can be executed synchronously or sequentially; this embodiment does not specifically limit this. By building agents with different functions and pre-assigning different retrieval tasks to multiple agents, parallel retrieval of different types of customer data is achieved during the customer data supplementation process, further improving the efficiency and accuracy of data retrieval.

[0071] Furthermore, after finding trade-related data of potential customers, the intelligent agent can input it into the customer management platform.

[0072] Specifically, an AI search component can be pre-created in the customer management platform. Once triggered, the platform's current display will switch to the agent's search results interface for trade-related data. Specifically, the display interface corresponding to the AI ​​search component maintains different types of fields, each representing different types of trade-related data, such as... Figure 1b As shown, the AI ​​search component's display interface includes multiple search fields such as "Company Information Search," "Contact Information Search," and "Product Import / Export Information." Each field's display bar contains multiple sub-fields. For example, the "Company Information Search" field's display box includes fields like "Customer Name," "Website," "Email," "Phone Number," and "Action." The "Contact Information Search" field's display box includes sub-fields like "Name," "Job Title," "Email," "Phone Number," and "Action." The "Import / Export Information Search" field's display box includes sub-fields like "Product," "Date," "Company Name," "Code," and "Action." The Agent can write the retrieved information into the corresponding display boxes according to the sub-fields. This setup facilitates sales personnel's viewing and editing of potential clients' trade-related data.

[0073] In this embodiment, after obtaining trade-related data from each potential customer, the Agent can extract key features relevant to business cooperation from this data, such as product order quantity, transaction amount, and / or import / export volume. Analyzing these key features can determine the cooperation value of each potential customer.

[0074] Understandably, in order for the Agent to have the function of evaluating the cooperation value of potential customers, it needs to be configured with the following prompts: key feature information related to business cooperation to be extracted from trade-related data, the evaluation process and evaluation criteria for the cooperation value of each potential customer, etc. By training the Fourth Agent using the above prompts, the Fourth Agent can be equipped with the function of judging the cooperation value of customers.

[0075] As an optional implementation, after the fourth agent is built and trained, it can execute the evaluation process of the cooperation value of each potential customer by calling the interface of the built fourth agent, so as to identify high-value potential customers. That is, the fourth agent analyzes the trade-related data of the potential customer to obtain key features and their feature values ​​of different dimensions related to business cooperation; the fourth agent performs a weighted summation of multiple key feature values ​​to obtain the comprehensive score corresponding to the potential customer, wherein the comprehensive score is used to evaluate the cooperation value.

[0076] The key features related to business cooperation in different dimensions may include the following five dimensions, although more dimensions may be set. This embodiment does not specifically limit this: 1. Product transaction volume, such as import and export quantity, order volume, minimum order quantity, etc.; 2. Transaction amount of the product; 3. The degree of alignment with the current sales party's business; 4. Compatibility with the current seller's products; 5. Other criteria, such as the size of potential clients' companies, which can be reflected by information such as whether they have an official website or a detailed product list.

[0077] In this embodiment, after the fourth agent is invoked, for each potential customer, the fourth agent can analyze the trade-related data of the potential customer to obtain the product transaction volume information and product transaction amount, and can determine the score corresponding to the product transaction volume information and the score corresponding to the product transaction amount.

[0078] Furthermore, by analyzing trade-related data of potential clients, the fourth agent can determine the client's main business or main products, and assess the matching degree between this main business and the previous seller's main business, as well as the matching degree between the potential client's main products and the current seller's products. For example, it can determine whether the potential client's main products match the current seller's products based on product functionality, usability, raw materials, etc., or whether the potential client's main products are upstream or downstream products of the current seller's products. The more successful matches during the matching process, the higher the matching degree.

[0079] Furthermore, by analyzing the potential client's trade-related data, the fourth agent can determine the potential client's company size. For example, if the potential client has an official website and / or a detailed product list, it indicates that the potential client's company size meets the requirements, and a preset score, such as 1, can be assigned to this feature. If the potential client does not have an official website and does not have a detailed product list, it indicates that the potential client's company size does not meet the requirements, and another preset score, such as 0, can be assigned to this feature.

[0080] In this embodiment, during the evaluation of potential customers, the fourth agent can perform a weighted summation of multiple key feature values ​​to obtain a comprehensive score for each potential customer. The higher the comprehensive score, the higher the cooperation value of the potential customer. The weight information corresponding to each key feature value can be set based on the needs of the sales party in the actual application scenario. The sales party can assign a higher weight value to the feature that it values ​​more.

[0081] In this embodiment, by assessing the business cooperation value of potential customers, high-value and highly matched customers can be screened from a large pool of potential clients. This allows for targeted follow-up with each high-value and highly matched customer, avoiding resource waste and ineffective investment. It not only reduces cooperation risks but also accurately meets customer needs, increasing the success rate of cooperation. Using an intelligent agent to evaluate the cooperation value of potential customers instead of sales personnel not only solves the problem of low efficiency in manual operations but also effectively ensures customer quality, preventing delays in follow-up due to inaccurate identification of high-value customers. Especially when dealing with a large number of potential customers, it effectively improves the accuracy and efficiency of customer cooperation value evaluation, thereby enhancing the company's operational efficiency.

[0082] S120. Select potential customers whose cooperation value meets the requirements as potential customers to be explored, generate product promotion content that matches each potential customer, and send the product promotion content to the corresponding potential customer.

[0083] In this embodiment, whether the cooperation value meets the requirements can be set according to the needs of the actual application scenario. For example, potential customers whose key feature values ​​in each dimension all reach the corresponding preset threshold can be regarded as potential customers to be explored. Alternatively, potential customers whose comprehensive score reaches the comprehensive score threshold can be regarded as potential customers to be explored. Or, the seller can directly specify potential customers to be explored from potential customers based on the feature values ​​of key features in each dimension. This embodiment does not specify the criteria for judging potential customers.

[0084] In this embodiment, the Agent can be invoked to identify potential customers from multiple potential customers and generate product promotion content that matches each potential customer.

[0085] Understandably, to enable the Agent to customize product promotion content for potential customers, the Agent (referred to as Agent 5 here for differentiation) needs to be configured with the following prompts: rules for identifying potential customers, rules and processes for generating product promotion content, product images, product descriptions, and email templates for generating product promotion content. By training Agent 5 using these prompts, it gains the ability to identify potential customers and customize product promotion content for them. Furthermore, during Agent 5's training, an email address validity verification tool can be configured. Agent 5 can use this tool's API to verify the validity of potential customer email addresses.

[0086] Once the fifth agent is set up, it is invoked. This fifth agent can then personalize product promotion content according to the process defined during its training. The specific process includes the following steps A to B: A. Identify the validity of each potential customer's email address. For each potential customer with a valid email address, determine the types of products that the potential customer is interested in based on the customer's trade-related data, and obtain product images.

[0087] The fifth agent can call the interface of an email address validity verification tool to identify the validity of email addresses of potential clients. Next, the fifth agent can analyze the trade-related data of each potential client to determine the product types whose traded quantities, amounts, and import / export volumes reach corresponding set thresholds. This product type can then be identified as the type of product the potential client is interested in, and the fifth agent can retrieve images of this product type from the database.

[0088] B. Select the target template corresponding to the product type from multiple pre-configured email templates, and add relevant data about the product that the customer to be mined is interested in to obtain product promotion content.

[0089] The product promotion content includes key information such as product categories, product introductions, and product images. Different product types require different email templates. Furthermore, to enhance the human-like feel of the email templates, each template can be pre-edited and optimized manually. During the invocation of the Fifth Agent, the agent selects the target template corresponding to the product type from a pre-configured set of email templates. It can also retrieve relevant data from the product knowledge base that might interest the target customer, such as product advantage documents and product images, and add them to the target template to generate the product promotion content.

[0090] In this embodiment, by using agents to replace sales personnel in customizing product promotion content for each potential customer, enterprises can achieve efficient and precise product promotion, reducing labor and time costs and improving operational efficiency.

[0091] Furthermore, to enhance the personalization of product promotion content tailored for potential clients by the Agent, various personalization adjustment parameters can be set during the setup of the fifth Agent, and the parameter values ​​for each parameter can be configured to maximize the personalization of the fifth Agent's output. These personalization adjustment parameters include a first parameter, a second parameter, and a third parameter. The first parameter adjusts the level of detail in the output information, the second parameter adjusts the randomness of the output information, and the third parameter adjusts the selection range of product-related data in the product promotion content. By setting these multiple personalization adjustment parameters and continuously testing and configuring their values, the combined effect of these parameters during the fifth Agent's invocation ensures that the product promotion content generated by the fifth Agent is highly personalized. This allows potential clients to perceive the product promotion content as tailor-made for them, thereby increasing their willingness to cooperate and improving the success rate of cooperation.

[0092] In this embodiment, after the fifth agent customizes product promotion content for the target customers, it can send the generated product promotion content to the corresponding target customers via email. Specifically, when editing the email content, it can be edited using the target customer's commonly used language and another auxiliary language (e.g., English), and the information to be emphasized in the email (e.g., country information, product name, unique advantages and features of the product, etc.) can be bolded or highlighted. The target customer's commonly used language can be determined through their trade-related data. Specifically, a correspondence table between various countries and their corresponding commonly used languages ​​can be pre-established. When editing the email content, the agent can determine the country information of each target customer based on the target customer's trade-related data, and can determine the corresponding commonly used language based on the pre-established country-language correspondence table, thus avoiding the problem that the agent's large model cannot automatically determine the corresponding language based on the country information.

[0093] It should be noted that when the Fifth Agent sends emails to potential clients, it can use preset email addresses provided by a pre-defined email provider for product marketing. Using these preset email addresses not only avoids consuming excessive mailbox storage resources when sending large volumes of emails, but also prevents emails from being blocked by the email service provider's anti-spam policies, thus avoiding the problem of emails failing to be sent.

[0094] Furthermore, the customer information management interface of the customer management platform is pre-configured with fields indicating whether the email addresses of potential customers are valid and fields indicating the email sending status. The specific values ​​for the email validity field can include: valid (the email address exists and can receive emails), invalid (the email address does not exist or cannot receive emails), unable to verify (the server is busy or blocking verification requests), and incorrect email format (the entered email address does not conform to the standard format), etc. The specific values ​​for the email sending status field can be sent, not sent, etc. Sales personnel can clearly understand the follow-up status of each potential customer through the values ​​corresponding to these fields. For example, Figure 1c This invention provides a schematic diagram of the product promotion status of a customer management platform as shown in Embodiment 1 of the present invention. Figure 1c As shown, the display interface includes fields such as customer name, customer type, product type, completed customer, email address, AI email sending status, and email address validity. For the "email address validity" field, the Agent can record the verification result of each potential customer's email address in this field. For the "AI email sending status" field, the Agent can record the email sending result of each potential customer in this AI email sending status field. The advantage of this setup is that sales personnel can easily view the follow-up status of each potential customer at any time.

[0095] Furthermore, after the fifth agent sends emails promoting the product, it can regularly track the email feedback from potential customers.

[0096] S130. Identify feedback from each potential customer regarding the corresponding product promotion content to determine the cooperation intention of each potential customer.

[0097] As an optional implementation, the backend service can periodically identify the receipt status of preset sending email addresses to track the feedback of each potential customer regarding product promotion content; upon receiving feedback from a potential customer, the service can call an NLP service to perform semantic recognition on the feedback content to determine the potential customer's cooperation intention.

[0098] As another optional implementation, the Agent can be invoked to perform semantic recognition on the feedback from customers regarding product promotion content, and determine the cooperation intentions of the customers to be mined.

[0099] Understandably, to enable the Agent to determine the cooperation intentions of potential clients, the Agent (referred to as the Sixth Agent here to distinguish it from other Agents) needs to be configured with the following prompts: rules for judging automatic reply emails (e.g., identifying automatic reply emails based on email titles), criteria for classifying the degree of intent, the cooperation intention determination process, and historical email interaction cases between the current salesperson and its partners. These historical email interaction cases include historical intent tags. By training the Sixth Agent using these prompts, it gains the ability to identify feedback from potential clients and determine their cooperation intentions.

[0100] For example, after the sixth agent is built, it is invoked. This sixth agent can first filter emails with automatic replies and then identify the content of emails without automatic replies to obtain the cooperation intention level of each potential customer. The process for determining the cooperation intention of each potential customer by the sixth agent can be as follows: First, break down the customer's reply content and extract different types of keywords, such as demand-related keywords (e.g., application scenarios, product order volume, or product quantity), sentiment-related keywords (e.g., the product is excellent, we don't need this type of product), and decision-related keywords (e.g., price, quotation, discount, or cooperation process). Then, based on historical email interactions with the potential partner, the corresponding intention level is determined. For example, the intention level can be divided into five levels, from low to high: no intention, slightly intentional, moderate intention, high intention, and extremely high intention. The intention level can then be determined based on the number of demand-related keywords, sentiment-related keywords, and decision-related keywords. The more demand-related keywords, the more positive sentiment-related keywords, and / or the more decision-related keywords, the higher the intention level.

[0101] In this embodiment, by using an agent to replace manual identification of the feedback content of each potential customer, the degree of cooperation intention of each potential customer is determined. This solves the problem of time and energy limitations in manually reviewing email content one by one. Especially when the number of potential customers is large, it greatly improves the efficiency of email content processing, not only reduces costs, but also ensures the accuracy of intent identification results, and avoids the loss of high-intent customers due to delays in manual processing.

[0102] Furthermore, after determining the cooperation intentions of each potential client, the sixth agent can record the specific level of cooperation intention in the corresponding interface of the client management platform. For example, Figure 1d A schematic diagram of an AI activity interface in a customer management platform is provided for Embodiment 1 of the present invention, such as... Figure 1d As shown, an "AI Development" module is pre-created in the customer management platform's display interface. This module can be configured with multiple components to showcase AI activities, including "AI Activities" (such as generating product promotion content through an agent, sending product push content through an agent, and determining the cooperation intent of potential customers through an agent), "Activity Record Content" (showing the specific steps performed by the agent), "Contact Person Name," "Level of Intent," "Email Address," "Activity Record Content Images," and "Activity Record Content Attachments." The sixth agent can write the determined cooperation intent level of each potential customer into the corresponding fields, allowing sales personnel to clearly understand the cooperation intent of each potential customer.

[0103] Furthermore, after receiving email feedback from potential customers, the received email can be forwarded to designated sales personnel so that they can develop further marketing strategies.

[0104] The solution provided in this embodiment uses an intelligent agent to analyze trade-related data of potential customers to determine their cooperation value. This not only solves the problem of low efficiency in manual data analysis by sales personnel, but also effectively identifies high-value customers, ensuring customer quality and avoiding delays in follow-up due to inaccurate identification of high-value customers. By using an intelligent agent to replace manual labor in customizing product promotion content for potential customers, not only are labor and time costs reduced, but efficient and precise promotion of products is also achieved. Furthermore, by using an intelligent agent to identify feedback from potential customers and determine the degree of cooperation intention of each potential customer, the problem of time and energy limitations in manually reviewing email content is solved, and the accuracy of customer cooperation intention assessment results is ensured. Especially when there is a large number of potential customers, it significantly improves the efficiency of email content processing, reduces costs, avoids the loss of high-intent customers due to delays in manual processing, and increases the success rate of cooperation with potential customers.

[0105] Example 2 Figure 2 This is a flowchart illustrating a product marketing method based on RPA, AI, and LLM for implementing agents, as provided in Embodiment 2 of this application. In this embodiment, agents with different functions can be pre-built, and the calling order of each agent can be set. Furthermore, the product marketing method provided in this embodiment can be encapsulated as a functional module, which provides a corresponding API (Application Programming Interface). By calling this API, the product marketing method provided in this embodiment can be implemented. Alternatively, each step of the product marketing method provided in this embodiment can be encapsulated into multiple functional units based on the functions to be implemented. For example, a functional unit for supplementing the introduction of potential customer companies, a functional unit for supplementing the contact information of potential customers, a functional unit for supplementing the import and export data of potential customers, a functional unit for determining the cooperation value of each potential customer, a functional unit for sending product promotion content, and a functional unit for identifying customer feedback information, etc. Each functional unit provides a corresponding API, and each functional unit can be called individually or in combination, depending on the needs of the actual application scenario. The caller of the functional module and the API corresponding to each functional unit can be RPA. For example, RPA can be configured to call one or more APIs periodically, or it can be manually triggered by maintenance personnel. This embodiment does not specifically limit this. Figure 2As shown, the method provided in this embodiment includes: S210. Obtain basic information about potential customers from the customer management platform.

[0106] The basic information includes the company name and its country of origin.

[0107] S220. Invoke the first agent to search for company introductions related to the basic information of potential customers on the first information retrieval platform and enter them into the customer management platform.

[0108] S230. Call the second agent to search for contact information related to the basic information of potential customers on the second information retrieval platform and enter it into the customer management platform.

[0109] S240. Call the third agent to search for product import and export information related to the basic information of potential customers on the third information retrieval platform through the third agent, and enter it into the customer management platform.

[0110] Steps S220, S230, and S240 can be executed synchronously or sequentially; this embodiment does not impose any specific limitations.

[0111] In this embodiment, the prompts for the first Agent, the second Agent, and the third Agent during the training process all include: the calling interface of the corresponding information retrieval platform, as well as the input parameters and output parameters of the information retrieval platform. The information retrieval platform corresponding to the first Agent is the first information retrieval platform, the information retrieval platform corresponding to the second Agent is the second information retrieval platform, and the information retrieval platform corresponding to the third Agent is the third information retrieval platform. Each information retrieval platform is used to query different types of trade-related data of potential customers.

[0112] S250. Call the fourth agent to analyze the trade-related data of each potential customer, obtain key features and their feature values ​​in different dimensions related to business cooperation, and use the fourth agent to perform a weighted summation of multiple key feature values ​​to obtain a comprehensive score for each potential customer, and enter the comprehensive score for each potential customer into the customer management platform.

[0113] The comprehensive score is used to evaluate the value of cooperation. The prompts used by the fourth agent during training include, but are not limited to: key features related to business cooperation that need to be extracted from trade-related data, the evaluation process and criteria for assessing the cooperation value of each potential customer, etc. By training the fourth agent using these prompts, it gains the ability to judge the value of customer cooperation.

[0114] In this embodiment, the specific implementation of steps S210 to S250 can be referred to the description of the above embodiment, and will not be repeated here.

[0115] S260. Invoke the fifth agent to identify the validity of each potential customer's email address. For each potential customer with a valid email address, use the fifth agent to process the customer's trade-related data, determine the product types that the potential customer is interested in, and obtain product images. Then, use the fifth agent to select the target template corresponding to the product type from multiple pre-configured email templates, add the product-related data that the potential customer is interested in to the target template, obtain product promotion content, and use the fifth agent to edit the product promotion content in the language commonly used by the potential customer, and send it to the corresponding potential customer in the form of an email.

[0116] During the training of the fifth agent, the cue words configured for this agent include, but are not limited to: rules for identifying customers to be mined, rules and processes for generating product promotional content, product images, product descriptions, and email templates for generating product promotional content. By training the fifth agent using these cue words, the fifth agent is equipped with the ability to identify customers to be mined and to customize product promotional content for these customers.

[0117] Furthermore, to enhance the personalization of the product promotion content customized for potential clients by the Fifth Agent, various personalization adjustment parameters are set during the Fifth Agent's development process. The parameter values ​​for each parameter are configured to maximize the personalization of the Fifth Agent's output. These personalization adjustment parameters include a first parameter, a second parameter, and a third parameter. The first parameter adjusts the level of detail in the output information, the second parameter adjusts the randomness of the output information, and the third parameter adjusts the selection range of product-related data in the product promotion content. By setting these multiple personalization adjustment parameters and continuously experimenting to configure their values, the Fifth Agent, when invoked, achieves maximum personalization of the generated product promotion content through the combined effect of these configured parameters. This allows potential clients to perceive the product promotion content as tailor-made for them, thereby increasing their willingness to cooperate and improving the success rate of cooperation.

[0118] S270. Call the sixth agent to filter emails with automatic replies. For emails without automatic replies, use the sixth agent to identify the feedback content of each potential customer, obtain the cooperation intention level of each potential customer, and enter the cooperation intention level of each potential customer into the customer management platform.

[0119] During the development of the sixth agent, the prompts configured for it include, but are not limited to: rules for identifying emails with automatic replies (e.g., identifying emails with automatic replies based on their subject lines), criteria for classifying levels of intent, a process for determining cooperation intent, and historical email interaction cases between the current salesperson and its partners. These historical email interaction cases include historical intent tags. By training this sixth agent, it gains the ability to identify feedback from potential clients regarding product promotion content and determine their cooperation intent.

[0120] In this embodiment, the specific implementation of steps S260 to S270 can be referred to the description of the above embodiment, and will not be repeated here.

[0121] In this embodiment, by pre-building agents with different functions and setting the calling order of each agent, the product marketing process can sequentially call each agent according to the calling order. Each agent performs its specific function, executing different steps in the product marketing process, thereby optimizing the execution accuracy and effectiveness of each stage of the product marketing process and achieving automated and efficient product marketing. Furthermore, by using multiple agents in combination, it is convenient to test and verify the agents at each stage during the agent setup process. This approach offers strong scalability and maintainability, adapting to changes in business scale.

[0122] Example 3 Figure 3 This application provides a structural block diagram of a product marketing device based on RPA, AI, and LLM to implement an agent, as shown in Embodiment 3 of this application. Figure 3 As shown, the device includes: a cooperation value determination module 310, a product promotion content generation module 320, a product promotion content sending module 330, and a cooperation intention determination module 340, wherein... The cooperation value determination module 310 is configured to extract key features related to business cooperation from the trade-related data of each potential customer, and determine the cooperation value of each potential customer based on the key features. The trade-related data includes company introduction, contact information and / or product import and export information, and the key features include product order quantity, transaction amount and / or import and export quantity. The product promotion content generation module 320 is configured to identify potential customers whose cooperation value meets the requirements as potential customers to be explored, and generate product promotion content that matches each potential customer. The product promotion content sending module 330 is configured to send product promotion content to the corresponding target customers; The cooperation intention determination module 340 is configured to identify feedback information from each potential customer regarding the corresponding product promotion content in order to determine the cooperation intention of each potential customer.

[0123] In one embodiment, the apparatus provided by the present invention further includes: The basic information acquisition module is configured to acquire basic information about potential customers from the customer management platform. This basic information includes the company name and country of origin. The trade-related data query module is configured to search for trade-related data of potential customers from multiple information query platforms based on basic information and enter it into the customer management platform.

[0124] In one implementation, potential customers include: Historical clients with initial cooperation intentions found on the customer management platform, and... Customers who have cooperative relationships with various competitors and / or customers who have transactions with competitors, as determined by competitor data and / or competitor product data. Accordingly, the apparatus provided in this embodiment of the invention further includes: Basic information on customers with whom you have cooperative relationships and customers with whom you have transactions with competing products will be entered into the customer management platform.

[0125] In one implementation, the trade-related data query module includes: The Agent invocation unit invokes multiple agents that have been set up to retrieve various types of trade-related data related to the basic information of each potential customer from different information query platforms. Each agent includes: The first agent is used to log in to the first information retrieval platform to search for company profiles related to the basic information of each potential customer. The second agent is used to log in to the second information retrieval platform to search for contact information related to the basic information of each potential customer. The third agent is used to log in to the third information retrieval platform to search for product import and export information related to the basic information of each potential customer.

[0126] In one implementation, the trade-related data query module includes: The RPA task execution unit is used to log in to different information query platforms via RPA to find various types of trade-related data related to the basic information of each potential customer.

[0127] In one implementation, the cooperation value determination module 310 is specifically configured as follows: The established fourth agent is invoked to analyze the trade-related data of each potential customer, thereby obtaining key features and their feature values ​​in different dimensions related to business cooperation. The fourth agent performs a weighted summation of multiple key feature values ​​to obtain a comprehensive score for the potential customer, which is used to evaluate the value of cooperation.

[0128] In one implementation, the product promotion content generation module 320 is specifically configured as follows: The established fifth agent is invoked to identify the validity of each customer email address to be mined; For each potential customer with a valid email address, the fifth agent analyzes the customer's trade-related data to determine the types of products the customer is interested in and obtains product images. The fifth agent selects the target template corresponding to the product type from multiple pre-configured email templates, and adds product-related data of interest to the customer to be mined into the target template to obtain product promotion content. Accordingly, the product promotion content sending module 330 is specifically configured as follows: The fifth agent edits the product promotion content into the language commonly used by the target customers and sends it to them via email. The language commonly used by the target customers is determined through trade-related data.

[0129] In one implementation, during the construction of the fifth agent, various types of personalized adjustment parameters are set for the fifth agent, and the parameter values ​​of each personalized adjustment parameter are configured to maximize the personalization of the output results of the fifth agent. The personalized adjustment parameters include a first parameter, a second parameter, and a third parameter. The first parameter is used to adjust the level of detail of the output information, the second parameter is used to adjust the randomness of the output information, and the third parameter is used to adjust the selection range of product-related data in the product promotion content.

[0130] In one implementation, the cooperation intention determination module 340 is configured to: Upon receiving a reply email from a customer to be identified, the established sixth agent is invoked to filter emails of the automatically replied type. For emails that are not automatically replied to, the sixth agent identifies the feedback from each potential client to determine their level of cooperation intention. The sixth agent is trained during configuration using the following prompts: The rules for judging automatic reply emails, the criteria for classifying the degree of intent, and historical email interaction cases between the current sales party and its partners.

[0131] In one implementation, the cooperation intention determination module is configured to: Periodically identify the receipt status of preset sending email addresses to track feedback from potential customers regarding product promotion content; Upon receiving feedback from any potential client, the system invokes Natural Language Processing (NLP) services to perform semantic recognition on the feedback content, thereby determining the client's cooperation intentions.

[0132] The functions of each module in the devices of this application embodiment can be found in the corresponding descriptions in the above methods, and will not be repeated here.

[0133] Example 4 Figure 4 This is a structural block diagram of an electronic device provided in Embodiment 4 of this application. Figure 4 As shown, the electronic device includes a memory 910 and a processor 920. The memory 910 stores a computer program that can run on the processor 920. When the processor 920 executes the computer program, it implements the product marketing method based on RPA, AI, and LLM in the above embodiments. The number of memories 910 and processors 920 can be one or more.

[0134] The electronic device also includes: The communication interface 930 is used to communicate with external devices and exchange and transmit data.

[0135] If the memory 910, processor 920, and communication interface 930 are implemented independently, they can be interconnected via a bus to communicate with each other. This bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. This bus can be divided into address bus, data bus, control bus, etc. For ease of representation, Figure 4The bus is represented by a single thick line, but this does not mean that there is only one bus or one type of bus.

[0136] Optionally, in a specific implementation, if the memory 910, processor 920, and communication interface 930 are integrated on a single chip, then the memory 910, processor 920, and communication interface 930 can communicate with each other through an internal interface.

[0137] This application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the method provided in this application.

[0138] This application also provides a chip, which includes a processor for calling and executing instructions stored in a memory, causing a communication device on which the chip is installed to perform the method provided in this application.

[0139] This application also provides a chip, including: an input interface, an output interface, a processor, and a memory. The input interface, output interface, processor, and memory are connected through an internal connection path. The processor is used to execute code in the memory. When the code is executed, the processor is used to execute the method provided in the application embodiment.

[0140] It should be understood that the aforementioned processor can be a Central Processing Unit (CPU), or other general-purpose processors, Digital Signal Processors (DSPs), Application Specific Integrated Circuits (ASICs), Field-Programmable Gate Arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. General-purpose processors can be microprocessors or any conventional processor. It is worth noting that the processor can be a processor supporting Advanced Reduced Instruction Set Machines (ARM) architecture.

[0141] Further, optionally, the aforementioned memory may include read-only memory and random access memory, and may also include non-volatile random access memory. The memory may be volatile or non-volatile, or may include both. Non-volatile memory may include read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically erasable programmable read-only memory (EEPROM), or flash memory. Volatile memory may include random access memory (RAM), which serves as an external cache. Many forms of RAM are available by way of example, but not limitation. Examples include Static Random Access Memory (SRAM), Dynamic Random Access Memory (DRAM), Synchronous DRAM (SDRAM), Double Data Rate Synchronous DRAM (DDR SDRAM), Enhanced Synchronous DRAM (ESDRAM), Synchlink DRAM (SLDRAM), and Direct Rambus RAM (DR RAM).

[0142] In the above embodiments, implementation can be achieved, in whole or in part, through software, hardware, firmware, or any combination thereof. When implemented in software, it can be implemented, in whole or in part, as a computer program product. A computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, all or part of the flow or function according to this application is generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transferred from one computer-readable storage medium to another.

[0143] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of this application. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of those different embodiments or examples.

[0144] Furthermore, the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features indicated. Thus, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, "a plurality of" means two or more, unless otherwise explicitly specified.

[0145] Any process or method description in the flowchart or otherwise herein can be understood as representing a module, segment, or portion of code comprising one or more executable instructions for implementing a particular logical function or process. Furthermore, the scope of the preferred embodiments of this application includes additional implementations in which functions may be performed not in the order shown or discussed, including substantially simultaneously or in reverse order depending on the functionality involved.

[0146] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a sequenced list of executable instructions for implementing logical functions, and can be embodied in any computer-readable medium for use by, or in conjunction with, an instruction execution system, apparatus or device (such as a computer-based system, a processor-included system or other system that can fetch and execute instructions from, an instruction execution system, apparatus or device).

[0147] It should be understood that various parts of this application can be implemented using hardware, software, firmware, or a combination thereof. In the above embodiments, multiple steps or methods can be implemented using software or firmware stored in memory and executed by a suitable instruction execution system. All or part of the steps of the methods in the above embodiments can be implemented by a program instructing related hardware, the program being stored in a computer-readable storage medium, which, when executed, includes one or a combination of the steps of the method embodiments.

[0148] Furthermore, the functional units in the various embodiments of this application can be integrated into a processing module, or each unit can exist physically separately, or two or more units can be integrated into a module. The integrated module can be implemented in hardware or as a software functional module. If the integrated module is implemented as a software functional module and sold or used as an independent product, it can also be stored in a computer-readable storage medium. This storage medium can be a read-only memory, a disk, or an optical disk, etc.

[0149] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any person skilled in the art can easily conceive of various variations or substitutions within the technical scope disclosed in this application, and these should all 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 product marketing method based on Robotic Process Automation (RPA), Artificial Intelligence (AI), and Large Language Modeling (LLM) to realize an intelligent agent, characterized in that, include: Key features related to business cooperation are extracted from the trade-related data of each potential customer, and the cooperation value of each potential customer is determined based on the key features. The trade-related data includes company introduction, contact information and / or product import and export information, and the key features include product order quantity, transaction amount and / or import and export quantity. Potential customers whose cooperation value meets the requirements are identified as potential customers to be explored, and product promotion content is generated to match each potential customer. Send product promotional content to the relevant potential customers; Identify feedback from potential clients regarding the promotional content for the corresponding products to determine their willingness to cooperate.

2. The method according to claim 1, characterized in that, The method further includes: Obtain basic information about potential customers from the customer management platform, including company name and country of origin; Based on the aforementioned basic information, trade-related data of each potential customer is retrieved from multiple information query platforms, and the trade-related data is entered into the customer management platform.

3. The method according to claim 1 or 2, characterized in that: The potential customers include: Historical clients with initial cooperation intentions found from the customer management platform, and... Customers who have cooperative relationships with various competitors and / or customers who have transactions with competitors, as determined by competitor data and / or competitor product data. Accordingly, the method further includes: Customer basic information on customers with whom you have cooperative relationships and customers with whom you have transactions with competing products will be entered into the customer management platform.

4. The method according to claim 2, characterized in that, The process of retrieving trade-related data for each potential customer from multiple information query platforms based on the aforementioned basic information includes: The system invokes multiple agents that have been set up to retrieve various types of trade-related data related to the basic information of each potential customer from different information query platforms. Each agent includes: First Agent is used to search for company profiles related to the basic information of each potential customer on the first information retrieval platform; The second agent is used to search for contact information related to the basic information of each potential customer on the second information retrieval platform. The third agent is used to search for product import and export information related to the basic information of each potential customer on a third information retrieval platform.

5. The method according to claim 2, characterized in that, The process of retrieving trade-related data for each potential customer from multiple information query platforms based on the aforementioned basic information includes: By logging into different information query platforms through RPA, various types of trade-related data related to the basic information of each potential customer can be found.

6. The method according to claim 1, characterized in that, The process of extracting key features related to business cooperation from the trade-related data of each potential customer, and determining the cooperation value of each potential customer based on these key features, includes: The fourth agent, which has been built, is invoked to analyze the trade-related data of each potential customer and obtain key features and their feature values ​​in different dimensions related to business cooperation. The fourth agent performs a weighted summation of multiple key feature values ​​to obtain a comprehensive score for each potential customer, which is used to evaluate the value of the cooperation.

7. The method according to claim 1, characterized in that, The generation of product marketing content that matches each target customer includes: The established fifth agent is invoked to identify the validity of each customer email address to be mined. For each potential customer with a valid email address, the fifth agent analyzes the customer's trade-related data to determine the types of products the customer is interested in and obtains product images. The fifth agent selects a target template corresponding to the product type from multiple pre-configured email templates, and adds product-related data of interest to the target customer to obtain product promotion content. Accordingly, sending product promotional content to the corresponding potential customers includes: The fifth agent edits the product promotion content in the language commonly used by the target customer and sends it to the target customer via email. The common language of the target customer is determined by the trade-related data.

8. The method according to claim 7, characterized in that, During the setup of the fifth agent, various types of personalized adjustment parameters are set for the fifth agent, and the parameter values ​​of each personalized adjustment parameter are configured to maximize the personalization of the fifth agent's output results. The personalized adjustment parameters include: a first parameter, a second parameter, and a third parameter. The first parameter is used to adjust the level of detail in the output information, the second parameter is used to adjust the randomness of the output information, and the third parameter is used to adjust the selection range of product-related data in the product promotion content.

9. The method according to claim 1, characterized in that, The process of identifying feedback from each potential customer regarding the corresponding product promotion content to determine their cooperation intentions includes: Upon receiving a reply email from a customer to be identified, the established sixth agent is invoked to filter emails of the automatically replied type. For emails that are not automatically replied to, the sixth agent identifies the feedback content of each potential customer to determine their level of cooperation intention. The sixth agent is trained during configuration using the following prompts: The rules for judging automatic reply emails, the criteria for classifying the degree of intent, and historical email interaction cases between the current sales party and its partners.

10. The method according to claim 1, characterized in that, The process of identifying feedback from each potential customer regarding the corresponding product promotion content to determine their cooperation intentions includes: Periodically identify the receipt status of preset sending email addresses to track feedback from potential customers regarding product promotion content; Upon receiving feedback from any potential client, the system invokes Natural Language Processing (NLP) services to perform semantic recognition on the feedback content, thereby determining the client's cooperation intentions.

11. A product marketing device based on Robotic Process Automation (RPA), Artificial Intelligence (AI), and Large Language Modeling (LLM) to realize an intelligent agent, characterized in that, include: The cooperation value determination module is configured to extract key features related to business cooperation from the trade-related data of each potential customer, and determine the cooperation value of each potential customer based on the key features. The trade-related data includes company introduction, contact information and / or product import and export information, and the key features include product order quantity, transaction amount and / or import and export quantity. The product promotion content generation module is configured to identify potential customers whose cooperation value meets the requirements as potential customers to be explored, and generate product promotion content that matches each potential customer. The product promotion content sending module is configured to send product promotion content to the corresponding potential customers; The cooperation intention determination module is configured to identify feedback from each potential customer regarding the corresponding product promotion content in order to determine the cooperation intention of each potential customer.

12. The apparatus according to claim 11, characterized in that, The device further includes: The basic information acquisition module is configured to acquire basic information of potential customers from the customer management platform, including company name and country of origin. The trade-related data query module is configured to search for trade-related data of potential customers from multiple information query platforms based on the basic information, and then enter the data into the customer management platform.

13. The apparatus according to claim 12, characterized in that, The trade-related data query module includes: The Agent invocation unit invokes multiple agents that have been set up to retrieve various types of trade-related data related to the basic information of each potential customer from different information query platforms. Each agent includes: The first agent is used to log in to the first information retrieval platform to search for company profiles related to the basic information of each potential customer. The second agent is used to log in to the second information retrieval platform to search for contact information related to the basic information of each potential customer. The third agent is used to log in to the third information retrieval platform to search for product import and export information related to the basic information of each potential customer.

14. The apparatus according to claim 11, characterized in that, The cooperation value determination module is specifically configured as follows: The established fourth agent is invoked to analyze the trade-related data of each potential customer, thereby obtaining key features and their feature values ​​in different dimensions related to business cooperation. The fourth agent performs a weighted summation of multiple key feature values ​​to obtain a comprehensive score for the potential customer, which is used to evaluate the value of the cooperation.

15. The apparatus according to claim 11, characterized in that, The product promotion content generation module is specifically configured as follows: The established fifth agent is invoked to identify the validity of each customer email address to be mined; For each potential customer with a valid email address, the fifth agent analyzes the customer's trade-related data to determine the types of products the customer is interested in and obtains product images. The fifth agent selects the target template corresponding to the product type from multiple pre-configured email templates, and adds product-related data of interest to the customer to be mined into the target template to obtain product promotion content. Accordingly, the product promotion content sending module is specifically configured as follows: The fifth agent edits the product promotion content in the language commonly used by the target customer and sends it to the target customer via email. The common language of the target customer is determined by the trade-related data.

16. An electronic device, characterized in that, include: A processor and a memory, wherein instructions are stored in the memory and loaded and executed by the processor to implement the method as claimed in any one of claims 1 to 10.

17. A computer-readable storage medium storing a computer program that, when executed by a processor, implements the method as described in any one of claims 1-10.

Images