Large language model-based sales plan generation method and device

Abstract

The invention discloses a big language model-based sales plan generation method and device, and the method comprises the steps: constructing a marketing knowledge base which comprises a commodity basic information database, a social media database, a competition pattern database and a high-quality sales plan reference database; performing fine tuning training on a large language model according to the high-quality sales plan reference database to obtain a sales plan generation model; and generating a target sales plan according to the marketing knowledge base and the sales plan generation model. According to the method, the marketing plan can be generated by finely adjusting the large language model and utilizing a plurality of databases, so that the efficiency and the applicability are improved, and the cost is reduced. The method can be widely applied to the technical field of artificial intelligence.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a method and device for generating a marketing plan based on a large language model. Background Art

[0002] To better achieve specific marketing goals, companies typically develop marketing plans. Traditional marketing plan development methods rely on manual processing, requiring systematic analysis of the market environment and consumer demand, integrating factors such as product, price, channel, and promotion, and integrating a large amount of statistical information. This process is time-consuming, labor-intensive, costly, and inefficient. Furthermore, traditional methods based on large language models are general-purpose and have poor results for highly specialized and creative marketing plans. The generated content is often difficult to use in real-world scenarios, resulting in low applicability.

[0003] In summary, the technical problems existing in the relevant technologies need to be improved. Summary of the Invention

[0004] The embodiments of the present invention provide a method and device for generating a marketing plan based on a large language model, which effectively improves efficiency and applicability and reduces costs.

[0005] In one aspect, an embodiment of the present invention provides a method for generating a marketing plan based on a large language model, comprising the following steps:

[0006] Build a marketing knowledge base, which includes a product basic information database, a social media database, a competition landscape database, and a high-quality marketing plan reference database;

[0007] Fine-tune the large language model based on the high-quality marketing plan reference database to obtain a marketing plan generation model;

[0008] A target marketing plan is generated based on the marketing knowledge base and the marketing plan generation model.

[0009] In some embodiments, when the marketing knowledge base is the commodity basic information database, constructing the marketing knowledge base includes:

[0010] Obtain product-related information from the e-commerce platform, wherein the product-related information includes one or more information of brand, product type, product name, price, and launch time;

[0011] The commodity basic information database is constructed based on the product related information.

[0012] In some embodiments, when the marketing knowledge base is the social media database, constructing the marketing knowledge base includes:

[0013] Obtain user reviews of products from social media platforms;

[0014] Using a large language model to structure the evaluation information to obtain product evaluation structured information;

[0015] Based on the commodity basic information database, the product evaluation structured information is standardized using a large language model to obtain standardized product evaluation information, wherein the standardized product evaluation information includes one or more of standardized brand information, standardized product type information, and standardized product name information;

[0016] The social media database is constructed based on the product evaluation structured information and the product evaluation standardized information.

[0017] In some embodiments, the method of standardizing the product evaluation structured information using a large language model based on the commodity basic information database to obtain standardized product evaluation information includes:

[0018] Using the product review target information and all brands in the commodity basic information database as prompt words, and using a large language model to perform brand matching to obtain the brand standardized information, the product review target information includes one or more information of the brand, product type, and product name in the product review structured information;

[0019] Using the product evaluation target information and all product types corresponding to the brand standardized information in the commodity basic information database as prompt words, and using a large language model to perform product type matching to obtain the product type standardized information;

[0020] All product names corresponding to the product evaluation target information and the product type standardized information in the commodity basic information database are used as prompt words, and product names are matched using a large language model to obtain the standardized product name information.

[0021] In some embodiments, when the marketing knowledge base is the competition landscape database, constructing the marketing knowledge base includes:

[0022] Obtain information on the industry's competitive landscape from industry product research reports and online press releases;

[0023] Extracting competitive product information corresponding to each brand in the commodity basic information database from the industry competition landscape information, the competitive product information including one or more of leading competitors, emerging challengers, substitutes, cross-border competitors, market share, market share of segmented markets, and sales data of e-commerce channels;

[0024] The competition landscape database is constructed based on the competitor information.

[0025] In some embodiments, when the marketing knowledge base is the high-quality marketing plan reference database, constructing the marketing knowledge base includes:

[0026] Obtain high-quality marketing plans;

[0027] Utilize a large language model to perform basic information analysis on the high-quality marketing plan to obtain plan-related information, including one or more of brand, product type, product name, potential selling points of the product, potential product deficiencies, core competitiveness, competitor status, macro-environmental status, market value, current marketing actions of competitors, and marketing goals;

[0028] Structuring the information related to the plan to obtain structured information of the plan;

[0029] A high-quality marketing plan reference database is constructed based on the high-quality marketing plan and the plan structured information.

[0030] In some embodiments, fine-tuning the large language model based on the high-quality marketing plan reference database to obtain a marketing plan generation model includes:

[0031] Using the structured information of the marketing plan in the high-quality marketing plan reference database as model input;

[0032] Outputting the high-quality marketing plan cases in the high-quality marketing plan case reference database as a model;

[0033] constructing a sample data set according to the model input and the model output;

[0034] The large language model is fine-tuned and trained using the sample data set to obtain the marketing plan generation model.

[0035] In some embodiments, generating a target marketing plan based on the marketing knowledge base and the marketing plan generation model includes:

[0036] Use the big model to analyze the needs of user input prompt words to obtain basic product information and planning requirements;

[0037] Standardizing the basic information of the commodity according to the commodity basic information database to obtain standardized commodity information;

[0038] Performing a first statistical analysis based on the standardized product information and the social media database to obtain product summary information;

[0039] Performing a second statistical analysis based on the standardized product information and the competition landscape database to obtain competitor situation analysis information;

[0040] Use the large language model to analyze the industry environment based on the industry's macro-policy information to obtain macro-environmental analysis information;

[0041] Use the large language model to evaluate and analyze the evaluation information of the current product to obtain the market value information of the current product;

[0042] Use the large language model to analyze the competitor's official current marketing actions and obtain information about the competitor's current marketing actions;

[0043] Extracting reference marketing plans from the high-quality marketing plan reference database based on the standardized product information, the product summary information, the competitor situation analysis information, the macro-environment situation analysis information, the market value information of the current product, the competitor's current marketing action information and marketing goals;

[0044] The commodity standardization information, the commodity summary information, the competitor situation analysis information, the macro-environment situation analysis information, the market value situation information of the current product, the competitor's current marketing action situation information, the marketing goals and the reference marketing plan are input into the marketing plan generation model to obtain the target marketing plan.

[0045] In some embodiments, performing a first statistical analysis based on the standardized product information and the social media database to obtain product summary information includes:

[0046] Extracting product-related information from the social media database based on the standardized product information;

[0047] Perform statistical type classification based on the purchase signal situation to obtain statistical type classification results, wherein the statistical type classification results include product demand, potential users, purchase intention and consumption;

[0048] Based on the product-related information and the statistical type classification results, statistics are collected on the evaluation of the product and brand in each statistical type, and word frequency statistics are performed on the keywords to obtain statistical results;

[0049] Based on the statistical results, a large language model is used to summarize the statistical results to obtain the product summary information, which includes potential selling points, potential shortcomings and core competitiveness of the product.

[0050] On the other hand, an embodiment of the present invention provides a marketing plan generation device based on a large language model, comprising:

[0051] The first module is used to build a marketing knowledge base, which includes a basic product information database, a social media database, a competition landscape database, and a high-quality marketing plan reference database;

[0052] The second module is used to fine-tune the large language model based on the high-quality marketing plan reference database to obtain a marketing plan generation model;

[0053] The third module is used to generate a target marketing plan based on the marketing knowledge base and the marketing plan generation model.

[0054] In another aspect, an embodiment of the present invention provides a computer device, comprising:

[0055] at least one processor;

[0056] at least one memory for storing at least one program;

[0057] When the at least one program is executed by the at least one processor, the at least one processor implements the method.

[0058] On the other hand, an embodiment of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described above is implemented.

[0059] The beneficial effects of the present invention are as follows:

[0060] The embodiment of the present invention first constructs a marketing knowledge base, then fine-tunes and trains a large language model based on a high-quality marketing plan reference database to obtain a marketing plan generation model, and then generates a target marketing plan based on the marketing knowledge base and the marketing plan generation model. In this way, marketing plan generation can be achieved by fine-tuning the large language model and utilizing multiple databases, thereby improving efficiency and applicability and reducing costs.

[0061] Other features and advantages of the present invention will be described in the following description, and in part will become apparent from the description, or will be understood by practicing the present invention. The purpose and other advantages of the present invention can be realized and obtained through the structures particularly pointed out in the description and the drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] In order to more clearly illustrate the technical solutions in the embodiments of the present application, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0063] Figure 1 This is a flowchart of a method for generating a marketing plan based on a large language model according to an embodiment of the present invention;

[0064] Figure 2 This is a structural diagram of a marketing plan generation device based on a large language model according to an embodiment of the present invention;

[0065] Figure 3 The figure is a schematic diagram of the hardware structure of a computer device according to an embodiment of the present invention. DETAILED DESCRIPTION

[0066] In order to make the purpose, technical solutions and advantages of the present application clearer, the present application is further described in detail below with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present application and are not intended to limit the present application. When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the embodiments of the present application. They are merely examples of devices and methods consistent with some aspects of the embodiments of the present application as detailed in the appended claims.

[0067] It will be understood that the terms "first", "second", etc. used in this application may be used herein to describe various concepts, but unless otherwise specified, these concepts are not limited by these terms. These terms are only used to distinguish one concept from another. For example, without departing from the scope of the embodiments of the present application, the first information may also be referred to as the second information, and similarly, the second information may also be referred to as the first information. Depending on the context, the words "if" and "if" as used herein may be interpreted as "at the time of" or "when" or "in response to determining".

[0068] The terms "at least one", "plurality", "each", "any", etc. used in this application include "at least one", "two" or more, "plurality" or "each", "any" or "any one", "each" or "any one" as used herein.

[0069] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the art to which this application pertains. The terms used herein are for the purpose of describing the embodiments of this application only and are not intended to limit this application.

[0070] Before explaining the embodiments of the present application in detail, some of the nouns and terms involved in the embodiments of the present application are first explained. The nouns and terms involved in the embodiments of the present application are subject to the following explanations.

[0071] Large Language Models (LLMs) are deep learning models trained using large amounts of text data, enabling them to generate natural language text or understand the meaning of text. These models can be trained on massive datasets to provide in-depth knowledge and language production on a variety of topics. Their core idea is to learn the patterns and structure of natural language through large-scale unsupervised training, simulating, to a certain extent, the human language cognition and generation process.

[0072] RAG (Retrieval-Augmented Generation): is a technology that generates more accurate and reliable content by combining retrieval technology and large language models (LLM).

[0073] In the prior art, companies typically develop marketing plans to better achieve specific marketing goals (such as increasing sales, expanding market share, and enhancing brand awareness). However, developing a marketing plan typically requires a systematic analysis of the market environment and consumer demand, integrating factors such as product, pricing, distribution channels, and promotional features. This process is time-consuming, labor-intensive, and costly, with low efficiency. Using AI technology, particularly large language models, to assist in generation can significantly reduce costs. High-performing large language models such as DeepSeek-R1 can produce good results when using high-quality prompts. However, without Retrieval-Augmented Generation (RAG), they often produce illusory results, resulting in content that is inconsistent with reality and unusable. Marketing plans must be based on a variety of information and utilize RAG technology. The core of RAG technology is how to find the most relevant information to assist in generation. Existing methods directly parse information from various sources and build a knowledge base. Then, based on prompts, they identify relevant information. This information and prompts are then fed into the large model to generate content. Since the development of marketing plans often requires extensive statistical information, and most large language models are poorly suited for mathematical statistics, this approach often produces poor quality results. Furthermore, existing large language models are generally general-purpose, making them ineffective for directly generating highly professional and creative marketing plans. The generated content often proves unsuitable for real-world scenarios, resulting in low applicability.

[0074] In view of this, this embodiment constructs a marketing knowledge base, fine-tunes the large language model, obtains a marketing plan generation model, and generates a target marketing plan based on the marketing knowledge base and the marketing plan generation model, combined with a large amount of statistical information, thereby improving efficiency and applicability and reducing costs.

[0075] The marketing plan generation method based on a large language model provided in the embodiment of the present application relates to the field of artificial intelligence technology. The marketing plan generation method based on a large language model provided in the embodiment of the present application can be applied to a terminal, can also be applied to a server, and can also be software running in a terminal or a server. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, a smart speaker, a smart watch, and a car terminal, etc., but is not limited to this; the server side can be configured as an independent physical server, or can be configured as a server cluster or distributed system composed of multiple physical servers, and can also be configured as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. The server can also be a node server in a blockchain network; the software can be an application that implements a marketing plan generation method based on a large language model, etc., but is not limited to the above forms.

[0076] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments in which tasks are performed by remote processing devices connected via a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0077] The following is a detailed explanation of the embodiments of the present application with reference to the accompanying drawings:

[0078] Figure 1 This is an optional flowchart of the method for generating a marketing plan based on a large language model provided in an embodiment of the present application. Figure 1 The method may include but is not limited to steps S101 to S103.

[0079] Step S101: Build a marketing knowledge base, which includes a product basic information database, a social media database, a competition landscape database, and a high-quality marketing plan reference database;

[0080] Step S102: fine-tune the large language model based on a high-quality marketing plan reference database to obtain a marketing plan generation model;

[0081] Step S103: Generate a target marketing plan based on the marketing knowledge base and the marketing plan generation model.

[0082] Steps S101 to S103 shown in the embodiment of the present application realize the generation of marketing plans, improve efficiency and applicability, and reduce costs.

[0083] In some embodiments, in step S101, when the marketing knowledge base is a commodity basic information database, constructing the marketing knowledge base may include but is not limited to the following steps:

[0084] Obtain product-related information from the e-commerce platform, including one or more of brand, product type, product name, price, and launch time;

[0085] Build a basic product information database based on product-related information.

[0086] In some embodiments, when constructing a basic commodity information database, product-related information can be first obtained from mainstream e-commerce platforms, where the product-related information includes one or more information of brand, product type, product name, price and launch time, so as to subsequently standardize comments and other information related to product names, brands, etc. and assist in the generation of planning proposals, and then construct a basic commodity information database based on the product-related information.

[0087] In some embodiments, in step S101, when the marketing knowledge base is a social media database, building the marketing knowledge base may include but is not limited to steps S201 to S204:

[0088] Step S201: Obtain user evaluation information on the product from the social media platform;

[0089] Step S202: Using a large language model to structure the evaluation information to obtain product evaluation structured information;

[0090] Step S203: Based on the commodity basic information database, the structured product evaluation information is standardized using a large language model to obtain standardized product evaluation information. The standardized product evaluation information includes one or more of standardized brand information, standardized product type information, and standardized product name information.

[0091] Step S204: Construct a social media database based on the product evaluation structured information and the product evaluation standardized information.

[0092] In some embodiments, when building a social media database, you can first obtain user evaluation information on the product from mainstream social media platforms, and then use a large language model to structure the evaluation information to obtain product evaluation structured information for subsequent storage in the social media database. For example, product evaluation structured information may include brand, product type, product name, valid information inclusion (yes / no), purchase signal, overall emotional information (positive / neutral / negative), overall emotional segmentation information, keywords, product evaluation dimensions and brand evaluation dimensions. Purchase signals may include product demand, potential users, purchase intentions and consumption. Overall emotional segmentation information includes joy, trust, expectation, objectivity, thinking, suspicion, sadness, disgust, fear, anger and surprise. Brand evaluation dimensions may include brand reputation, product quality, customer service, innovation ability, market performance and user satisfaction. Furthermore, product evaluation dimensions can be designed specifically for different product types. For example, for cars, product evaluation dimensions may include: interior, pricing, maintenance cost, fuel consumption (endurance), space, comfort, control (driving experience), power, smart driving, safety and value retention rate.

[0093] After structuring the review information, since basic information such as brand and product name extracted by the large language model may be inconsistent, the large language model can be used to standardize the structured product review information based on the basic product information database to obtain standardized product review information. This ensures that the structured product review information is consistent with the information in the basic product information database. This standardized product review information includes one or more of standardized brand information, standardized product type information, and standardized product name information. Finally, a social media database is constructed based on the structured and standardized product review information.

[0094] In some embodiments, in step S203, based on the commodity basic information database, the product review structured information is standardized using a large language model to obtain standardized product review information, which may include but is not limited to the following steps:

[0095] The product review target information and all brands in the commodity basic information database are used as prompt words, and brand matching is performed using a large language model to obtain brand standardized information. The product review target information includes one or more of the brand, product type, and product name in the product review structured information.

[0096] The product evaluation target information and all product types corresponding to the brand standardized information in the commodity basic information database are used as prompt words, and the product type is matched using the large language model to obtain the standardized product type information;

[0097] All product names corresponding to the product evaluation target information and the standardized product type information in the commodity basic information database are used as prompt words, and the product names are matched using a large language model to obtain standardized product name information.

[0098] In some embodiments, based on the commodity basic information database, a large language model is used to standardize the product evaluation structured information to obtain standardized product evaluation information, wherein the standardized product evaluation information includes one or more of brand standardized information, product type standardized information, and product name standardized information. Standardization is performed in a step-by-step manner. The product evaluation target information and all brands in the commodity basic information database can be used as prompt words, and brand matching is performed using a large language model to obtain brand standardized information, wherein the product evaluation target information includes one or more of the brand, product type, and product name in the product evaluation structured information. For example, all brands in the commodity basic information database can be used as part of the prompt words for reference, and the extracted brand, product type, and product name can be used as part of the prompt words for the large language model to match the brands therein, determine the brand of the evaluation information, and obtain brand standardized information. Then, the product evaluation target information and all product types corresponding to the brand standardized information in the commodity basic information database are used as prompt words, and product type matching is performed using a large language model to obtain product type standardized information. Exemplarily, all product types of the brand in the commodity basic information database can be used as part of the prompt words for reference, and at the same time, the extracted brand, product type, and product name can be used as part of the prompt words to let the large language model match the product type of the review to obtain standardized product type information. Then, all product names corresponding to the product review target information and the standardized product type information in the commodity basic information database are used as prompt words, and the large language model is used to match the product names to obtain standardized product name information. Exemplarily, all product names under the brand and product type in the commodity basic information database can be used as part of the prompt words for reference, and at the same time, the extracted brand, product type, and product name can be used as part of the prompt words to let the large language model match the product names to obtain standardized product name information, so that the product types and product names can be gradually standardized, effectively reducing complexity and improving efficiency and accuracy.

[0099] In some embodiments, in step S101, when the marketing knowledge base is a competition landscape database, constructing the marketing knowledge base may include but is not limited to the following steps:

[0100] Obtain information on the industry's competitive landscape from industry product research reports and online press releases;

[0101] Extracting competitive product information corresponding to each brand in the product basic information database from industry competition information. Competitive product information includes one or more types of information such as leading competitors, emerging challengers, substitutes, cross-border competitors, market share, market share in different market segments, and sales data from e-commerce channels;

[0102] Build a competition landscape database based on competitor information.

[0103] In some embodiments, when constructing a competition landscape database, industry competition landscape information can be first obtained from industry product research reports and online press releases. Competitive product information corresponding to each brand in the basic product information database can then be extracted from this industry competition landscape information. Competitive product information includes one or more of the following: leading competitors, emerging challengers, substitutes, cross-border competitors, market share, market share in specific segments, and sales data from e-commerce channels. This information is then used to assist in the generation of planning proposals. The competition landscape database is then constructed based on this competitive product information.

[0104] In some embodiments, in step S101, when the marketing knowledge base is a high-quality marketing plan reference database, constructing the marketing knowledge base may include but is not limited to the following steps:

[0105] Obtain high-quality marketing plans;

[0106] Use the large language model to analyze the basic information of high-quality marketing plans to obtain relevant information about the plans, including one or more of the following: brand, product type, product name, potential selling points, potential product shortcomings, core competitiveness, competitor situation, macro-environment, market value, current marketing actions of competitors, and marketing goals;

[0107] Structural processing is performed on the relevant information of the planning case to obtain the structural information of the planning case;

[0108] Based on high-quality marketing plans and their structured information, a reference database of high-quality marketing plans is constructed.

[0109] In some embodiments, when constructing a high-quality marketing plan reference database, one may first obtain high-quality marketing plans, then use a large language model to perform basic information analysis on the high-quality marketing plans to obtain plan-related information, wherein the plan-related information includes one or more of the following: brand, product type, product name, potential selling points of the product, potential product deficiencies, core competitiveness, competitor situation, macro-environmental situation, market value situation, current marketing actions of competitors, and marketing goals. The plan-related information is then structured to obtain plan structured information. Finally, based on the high-quality marketing plans and the plan structured information, a high-quality marketing plan reference database is constructed to store the high-quality marketing plans and the plan structured information in the high-quality marketing plan reference database.

[0110] In some embodiments, in step S102, fine-tuning the large language model based on the high-quality marketing plan reference database to obtain a marketing plan generation model may include but is not limited to the following steps:

[0111] The structured information of marketing plans in the high-quality marketing plan reference database is used as the model input;

[0112] Output high-quality marketing plans from the high-quality marketing plan reference database as models;

[0113] Construct a sample data set based on model input and model output;

[0114] The sample data set is used to fine-tune the large language model to obtain a marketing plan generation model.

[0115] In some embodiments, relevant data is extracted from a high-quality marketing plan reference database for training a marketing plan generation model. The structured information of the plan in the high-quality marketing plan reference database can be used as the model input, i.e., prompt word X, and the high-quality marketing plan in the high-quality marketing plan reference database can be used as the model output, i.e., prompt word Y. Then, based on the model input and model output, a sample data set (X, Y) is constructed. For example, the prompt word of the model input can be: "Please generate a marketing plan based on the following information. 1. Basic information: xxx; 2. Potential selling points: xxx; 3. Core competitiveness: xxx; 4. Potential shortcomings: xxx; 5. Competitive situation: xxx; 6. Macro-environment situation: xxx; 7. Market value situation: xxx; 8. Competitor's latest marketing action situation: xxx; 9. Marketing goal: xxx." Finally, the sample data set is used to fine-tune the open source large language model to obtain a marketing plan generation model. For example, the large language model deepseek-r1 can be fine-tuned to obtain a marketing plan generation model to improve the performance of the model.

[0116] In some embodiments, in step S103, a target marketing plan is generated based on the marketing knowledge base and the marketing plan generation model, which may include but is not limited to steps S301 to S309:

[0117] Step S301: Use the big model to analyze the user input prompt words to obtain basic product information and planning requirements;

[0118] Step S302: Standardize the basic information of the product according to the product basic information database to obtain standardized product information;

[0119] Step S303: Perform a first statistical analysis based on the standardized product information and the social media database to obtain product summary information;

[0120] Step S304: Perform a second statistical analysis based on the product standardization information and the competition landscape database to obtain competitor analysis information.

[0121] Step S305: Use the large language model to perform industry environment analysis on the industry macro policy information to obtain macro environment situation analysis information;

[0122] Step S306: Use the large language model to evaluate and analyze the evaluation information of the current product to obtain the market value information of the current product;

[0123] Step S307: Use the large language model to analyze the competitor's official current marketing actions to obtain information about the competitor's current marketing actions;

[0124] Step S308: Extract reference marketing plans from a database of high-quality marketing plans based on product standardization information, product summary information, competitor analysis information, macro-environment analysis information, current product market value information, competitor current marketing action information, and marketing goals.

[0125] Step S309: Input the product standardization information, product summary information, competitor situation analysis information, macro-environment situation analysis information, current product market value information, competitor current marketing action information, marketing goals and reference marketing plan into the marketing plan generation model to obtain the target marketing plan.

[0126] In some embodiments, a large model can be used to first perform a demand analysis on the user input prompt words (such as "I want a marketing plan for Xiao M Automobile") to obtain basic product information and planning requirements (the default is to improve sales indicators). Then, based on the basic product information database, the basic product information is standardized to obtain standardized product information. Based on the standardized product information and the social media database, a first statistical analysis is performed to obtain product summary information. Based on the standardized product information and the competitive landscape database, a second statistical analysis is performed to obtain competitor situation analysis information. For example, competitor information can be obtained from the competitive landscape database based on the extracted standardized product information. The obtained product summary information can be summarized using a large language model to obtain the selling point information, core competitiveness, potential deficiencies, and other information of the competitor. At the same time, the market share, sales data, and other information of the competitor are obtained from the competitive landscape database and summarized using a large language model to obtain competitor situation analysis information.

[0127] Then, use crawler technology to retrieve macro-policy information about the product's industry online and use a large language model to analyze the industry environment. This provides information on the macro-environmental situation. Crawler technology can also be used to retrieve mainstream media reviews of the product online and analyze them using a large language model to obtain information on the product's market value. Crawler technology can also be used to crawl current official marketing activities from mainstream social media platforms and use a large language model to analyze competitors' official marketing activities to obtain information on their current marketing activities.

[0128] Finally, based on the standardized product information, product summary information, competitor analysis information, macro-environmental analysis information, current product market value information, competitor current marketing action information, and marketing objectives, relevant reference marketing plans are extracted from a database of high-quality marketing plans. The standardized product information, product summary information, competitor analysis information, macro-environmental analysis information, current product market value information, competitor current marketing action information, marketing objectives, and reference marketing plans are then input into a targeted, fine-tuned marketing plan generation model to generate a target marketing plan that meets the user's needs. For example, the prompt for inputting the marketing plan generation model could be: "Please generate a marketing plan based on the following basic information and reference marketing plans. Basic information is as follows: 1. Product information: xxx; 2. Potential selling points: xxx; 3. Core competitiveness: xxx; 4. Potential weaknesses: xxx; 5. Competitor situation: xxx; 6. Macro-environmental situation: xxx; 7. Market value: xxx; 8. Competitor's latest marketing action: xxx; 9. Marketing objectives: xxx. Reference marketing plans are as follows: xxx."

[0129] In some embodiments, in step S303, a first statistical analysis is performed based on the standardized product information and the social media database to obtain product summary information, which may include but is not limited to the following steps:

[0130] Extract product-related information from social media databases based on standardized product information;

[0131] Based on the purchase signal situation, statistical type classification is performed to obtain statistical type classification results, which include product demand, potential users, purchase intention and consumption;

[0132] Based on the product-related information and statistical type classification results, the evaluation of products and brands in each statistical type is counted separately, and the frequency of keywords is counted to obtain statistical results;

[0133] Based on the statistical results, a large language model is used to summarize the statistical results and obtain product summary information, which includes the product's potential selling points, potential product shortcomings, and core competitiveness.

[0134] In some embodiments, product-related information can be first extracted from a social media database based on standardized product information. Then, based on purchase signals, statistical type classification is performed to obtain statistical type classification results, wherein the statistical type classification results include product demand, potential users, purchase intention, and consumption. Then, based on the product-related information and statistical type classification results, evaluations of the product and brand in each statistical type are counted, and word frequency statistics are performed on keywords to obtain statistical results. Finally, based on the statistical results, the statistical results are summarized using a large language model to obtain product summary information. The product summary information includes potential selling points, potential product deficiencies, and core competitiveness. For example, product-related information can be retrieved from a social media database based on the extracted standardized product information, and statistical analysis is performed on the pre-processed structured information. Based on the purchase signals, the statistical results are divided into four categories: product demand, potential users, purchase intention, and consumption. Each category separately counts evaluations of various dimensions of the product and brand. At the same time, word frequency statistics are performed on the extracted keywords in each category to obtain statistical results. Based on the obtained statistical results, the statistical results are summarized using a large language model to obtain the product's potential selling points, potential product deficiencies, and core competitiveness. More generally, the second statistical analysis process is similar to the first statistical analysis process.

[0135] In some embodiments, this embodiment constructs a targeted marketing knowledge base based on the characteristics of the auxiliary information required during the development of marketing plans. This reduces the difficulty of analyzing information by the large model through information structuring, thereby improving the quality of the auxiliary information and making it more targeted and reference-oriented. This embodiment analyzes high-quality marketing plans and uses the marketing plans to generate sample data sets to fine-tune the existing open source general-purpose large language model, thereby improving the efficiency and applicability of the large language model in generating marketing plans.

[0136] The beneficial effects of implementing the embodiments of the present invention include: the embodiments of the present invention first construct a marketing knowledge base, then fine-tune and train the large language model based on a high-quality marketing plan reference database to obtain a marketing plan generation model, and then generate a target marketing plan based on the marketing knowledge base and the marketing plan generation model, thereby enabling the generation of marketing plans by fine-tuning the large language model and utilizing multiple databases, thereby improving efficiency and applicability and reducing costs.

[0137] like Figure 2 As shown, the embodiment of the present invention further provides a marketing plan generation device based on a large language model, comprising:

[0138] The first module 801 is used to build a marketing knowledge base, which includes a product basic information database, a social media database, a competition landscape database, and a high-quality marketing plan reference database;

[0139] The second module 802 is used to fine-tune the large language model based on the high-quality marketing plan reference database to obtain a marketing plan generation model;

[0140] The third module 803 is used to generate a target marketing plan based on the marketing knowledge base and the marketing plan generation model.

[0141] The contents of the above method embodiments are all applicable to the present device embodiments. The functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0142] like Figure 3 As shown, an embodiment of the present invention further provides a computer device, including:

[0143] at least one processor 901;

[0144] At least one memory 902, configured to store at least one program;

[0145] When at least one program is executed by at least one processor, the at least one processor implements Figure 1 The method shown.

[0146] The contents of the above method embodiments are all applicable to the present device embodiments. The functions specifically implemented by the present device embodiments are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0147] The embodiment of the present invention further provides a computer-readable storage medium, which stores a computer program, which is executed by a processor to implement Figure 1 The method shown.

[0148] The contents of the above method embodiments are all applicable to the present storage medium embodiment. The functions specifically implemented by the present storage medium embodiment are the same as those of the above method embodiments, and the beneficial effects achieved are also the same as those achieved by the above method embodiments.

[0149] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but are not intended to limit the scope of the present invention. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and essence of the present invention should be within the scope of the present invention.

Claims

1. A marketing plan generation method based on a large language model, characterized by: The following steps are involved: Build a marketing knowledge base, which includes a product basic information database, a social media database, a competition landscape database, and a high-quality marketing plan reference database; Fine-tune the large language model based on the high-quality marketing plan reference database to obtain a marketing plan generation model; A target marketing plan is generated based on the marketing knowledge base and the marketing plan generation model.

2. The method according to claim 1, characterized in that When the marketing knowledge base is the commodity basic information database, the step of constructing the marketing knowledge base includes: Obtain product-related information from the e-commerce platform, wherein the product-related information includes one or more information of brand, product type, product name, price, and launch time; The commodity basic information database is constructed based on the product related information.

3. The method according to claim 1, characterized in that When the marketing knowledge base is the social media database, the step of constructing the marketing knowledge base includes: Obtain user reviews of products from social media platforms; Using a large language model to structure the evaluation information to obtain product evaluation structured information; Based on the commodity basic information database, the product evaluation structured information is standardized using a large language model to obtain standardized product evaluation information, wherein the standardized product evaluation information includes one or more of standardized brand information, standardized product type information, and standardized product name information; The social media database is constructed based on the product evaluation structured information and the product evaluation standardized information.

4. The method according to claim 3, characterized in that The method of performing standardization processing on the product evaluation structured information using a large language model based on the commodity basic information database to obtain standardized product evaluation information includes: Using the product review target information and all brands in the commodity basic information database as prompt words, and using a large language model to perform brand matching to obtain the brand standardized information, the product review target information includes one or more information of the brand, product type, and product name in the product review structured information; Using the product evaluation target information and all product types corresponding to the brand standardized information in the commodity basic information database as prompt words, and using a large language model to perform product type matching to obtain the product type standardized information; All product names corresponding to the product evaluation target information and the product type standardized information in the commodity basic information database are used as prompt words, and product names are matched using a large language model to obtain the standardized product name information.

5. The method according to claim 1, wherein When the marketing knowledge base is the competition landscape database, the step of constructing the marketing knowledge base includes: Obtain information on the industry's competitive landscape from industry product research reports and online press releases; Extracting competitive product information corresponding to each brand in the commodity basic information database from the industry competition landscape information, the competitive product information including one or more of leading competitors, emerging challengers, substitutes, cross-border competitors, market share, market share of segmented markets, and sales data of e-commerce channels; The competition landscape database is constructed based on the competitor information.

6. The method according to claim 1, characterized in that When the marketing knowledge base is the high-quality marketing plan reference database, the step of constructing the marketing knowledge base includes: Obtain high-quality marketing plans; Utilize a large language model to perform basic information analysis on the high-quality marketing plan to obtain plan-related information, including one or more of brand, product type, product name, potential selling points of the product, potential product deficiencies, core competitiveness, competitor status, macro-environmental status, market value, current marketing actions of competitors, and marketing goals; Structuring the information related to the plan to obtain structured information of the plan; A high-quality marketing plan reference database is constructed based on the high-quality marketing plan and the plan structured information.

7. The method according to claim 1, characterized in that The method of fine-tuning the large language model based on the high-quality marketing plan reference database to obtain a marketing plan generation model includes: Using the structured information of the marketing plan in the high-quality marketing plan reference database as model input; Outputting the high-quality marketing plan cases in the high-quality marketing plan case reference database as a model; constructing a sample data set according to the model input and the model output; The large language model is fine-tuned and trained using the sample data set to obtain the marketing plan generation model.

8. The method according to claim 1, characterized in that Generating a target marketing plan based on the marketing knowledge base and the marketing plan generation model includes: Use the big model to analyze the needs of user input prompt words to obtain basic product information and planning requirements; Standardizing the basic information of the commodity according to the commodity basic information database to obtain standardized commodity information; Performing a first statistical analysis based on the standardized product information and the social media database to obtain product summary information; Performing a second statistical analysis based on the standardized product information and the competition landscape database to obtain competitor situation analysis information; Use the large language model to analyze the industry environment based on the industry's macro-policy information to obtain macro-environmental analysis information; Use the large language model to evaluate and analyze the evaluation information of the current product to obtain the market value information of the current product; Use the large language model to analyze the competitor's official current marketing actions and obtain information about the competitor's current marketing actions; Extracting reference marketing plans from the high-quality marketing plan reference database based on the standardized product information, the product summary information, the competitor situation analysis information, the macro-environment situation analysis information, the market value information of the current product, the competitor's current marketing action information and marketing goals; The commodity standardization information, the commodity summary information, the competitor situation analysis information, the macro-environment situation analysis information, the market value situation information of the current product, the competitor's current marketing action situation information, the marketing goals and the reference marketing plan are input into the marketing plan generation model to obtain the target marketing plan.

9. The method according to claim 8, characterized in that The first statistical analysis is performed based on the standardized product information and the social media database to obtain product summary information, including: Extracting product-related information from the social media database based on the standardized product information; Perform statistical type classification based on the purchase signal situation to obtain statistical type classification results, wherein the statistical type classification results include product demand, potential users, purchase intentions, and consumption; Based on the product-related information and the statistical type classification results, statistics are collected on the evaluation of the product and brand in each statistical type, and word frequency statistics are performed on the keywords to obtain statistical results; Based on the statistical results, a large language model is used to summarize the statistical results to obtain the product summary information, which includes potential selling points, potential shortcomings and core competitiveness of the product.

10. A marketing plan generation device based on a large language model, characterized in that: include: The first module is used to build a marketing knowledge base, which includes a basic product information database, a social media database, a competition landscape database, and a high-quality marketing plan reference database; The second module is used to fine-tune the large language model based on the high-quality marketing plan reference database to obtain a marketing plan generation model; The third module is used to generate a target marketing plan based on the marketing knowledge base and the marketing plan generation model.

Images