AI-driven marketing content automatic generation system, method and related device

Abstract

This application discloses an AI-driven automatic marketing content generation system, method, and related apparatus, relating to the field of marketing content generation. The system includes: a marketing data acquisition module for acquiring multi-source marketing analysis data of the target audience; a user profiling and demand mining module for segmenting user groups based on the multi-source marketing analysis data, mining the preferences and needs of each user group for marketing content, and determining the marketing theme direction; a marketing theme and content positioning module for determining the target marketing theme based on the marketing theme direction and market trend analysis results, and determining a marketing content positioning scheme in conjunction with marketing characteristics; a multimodal content generation module for generating multimodal marketing content based on the marketing content positioning scheme; and a compliance review and content optimization module for conducting compliance review and quality assessment of the multimodal marketing content. This application can automatically generate multimodal marketing content that meets the preferences and needs of user groups and market trends.

Description

Technical Field

[0001] This application relates to the field of marketing content generation, and in particular to an AI-driven automatic marketing content generation system, method, and related apparatus. Background Technology

[0002] Current traditional marketing content generation models are largely based on manual creation, supplemented by simple template editing tools. Marketing content creation lacks multi-source data support, relying solely on the experience of operations personnel for theme positioning and content writing. This results in low content alignment with market trends and user needs, and severe homogenization. While a few intelligent content generation tools have emerged, current intelligent solutions focus only on generating text-based copy, neglecting multimodal content collaborative creation. Furthermore, they rely on large-scale, high-quality training data, making model adaptability and implementation costs high for SMEs with limited data resources. Some intelligent marketing platforms also offer template-based content generation. Users can select industry templates, fill in key information, and the system automatically generates content. These tools achieve rapid content output through pre-set template libraries, but the fixed nature of templates leads to low content personalization and difficulty in dynamically adjusting to real-time market trends and user profiles. Summary of the Invention

[0003] The purpose of this application is to provide an AI-driven marketing content automatic generation system, method, and related apparatus that can automatically generate multimodal marketing content that meets the preferences and marketing trends of different user groups.

[0004] To achieve the above objectives, this application provides the following solution: Firstly, this application provides an AI-driven automatic marketing content generation system, comprising: The marketing data acquisition module is used to acquire multi-source marketing analysis data of the target audience; The user profiling and demand mining module is used to segment user groups based on user consumption behavior data and basic user information from multi-source marketing analysis data, using AI clustering algorithms. Based on the consumption behavior characteristics of each user group, AI association rule mining algorithms are used to mine the user group's preferences for marketing content. Based on the user group's preferences for marketing content, the marketing theme direction for the target audience is determined. The marketing theme and content positioning module is used to determine the target marketing theme of the target marketing object based on the marketing theme direction of the target marketing object and the marketing trend analysis results obtained based on multi-source marketing analysis data, and to determine the marketing content positioning plan based on the target marketing theme and marketing characteristics of the target marketing object. The multimodal content generation module is used to generate multimodal marketing content that meets the consistency verification requirements by applying AI natural language processing algorithms based on the marketing content positioning plan. The consistency verification is used to check whether the key points, tone, and style of the multimodal marketing content are consistent. The compliance review and content optimization module is used to review the compliance of multimodal marketing content with the dissemination platform rules data in the multimodal marketing analysis data. It modifies multimodal marketing content that fails the compliance review, and applies an AI marketing content quality assessment algorithm to conduct a quality assessment of preset dimensions for multimodal marketing content that passes the compliance review. It also optimizes and adjusts multimodal marketing content whose quality assessment results are lower than the preset quality score threshold.

[0005] Secondly, this application provides an AI-driven method for automatically generating marketing content, including: Obtain multi-source marketing analysis data from the target audience; Based on user consumption behavior data and basic user information from multi-source marketing analysis data, AI clustering algorithms are applied to segment user groups. Based on the consumption behavior characteristics of each user group, AI association rule mining algorithms are applied to mine the user group's preference needs for marketing content. Based on the user group's preference needs for marketing content, the marketing theme direction for the target audience is determined. The target marketing theme of the target audience is determined based on the marketing theme direction and the marketing trend analysis results obtained from multi-source marketing analysis data. The marketing content positioning plan is determined based on the target marketing theme and marketing characteristics of the target audience. Based on the marketing content positioning plan, AI natural language processing algorithms are used to generate multimodal marketing content that meets the consistency verification. The consistency verification is used to check whether the key points, tone, and style of the multimodal marketing content are consistent. Based on the dissemination platform rules data in the multimodal marketing analysis data, the dissemination platform compliance review of multimodal marketing content is conducted. Multimodal marketing content that fails the compliance review is modified. For multimodal marketing content that passes the compliance review, an AI marketing content quality assessment algorithm is applied to conduct a quality assessment in preset dimensions. Multimodal marketing content whose quality assessment results are lower than the preset quality score threshold is optimized and adjusted.

[0006] Thirdly, this application provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the above-described AI-driven automatic marketing content generation method.

[0007] Fourthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements the aforementioned AI-driven automatic marketing content generation method.

[0008] According to the specific embodiments provided in this application, this application has the following technical effects: This application provides an AI-driven automatic marketing content generation system, method, and related apparatus. This application can segment user groups, mine their preferences and needs for different user groups, and then generate corresponding multimodal marketing content based on marketing trend analysis results. Furthermore, the core expression points, communication tone, and expression style of the multimodal marketing content are consistent, and the various forms of marketing content are strongly correlated, forming a unified brand communication effect. This application can generate diverse marketing content, solving the problem of the single format of existing marketing content. Moreover, it does not require any preset fixed templates and can dynamically adjust marketing content based on real-time market trends and user profiles, resulting in a high degree of personalization. In addition, the automatic generation of marketing content in this application employs AI-driven clustering algorithms, association rule mining algorithms, natural language processing algorithms, and quality assessment algorithms, enabling a fully automated operation of marketing content generation based on AI algorithms. Attached Figure Description

[0009] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 A schematic diagram of functional modules of an AI-driven marketing content automatic generation system provided in an embodiment of this application; Figure 2 A schematic diagram of the specific functional units of a marketing data collection module provided in an embodiment of this application; Figure 3 A schematic diagram of the specific functional units of the user profiling and demand mining module provided in an embodiment of this application; Figure 4 A schematic diagram of the specific functional units of the marketing theme and content positioning module provided in an embodiment of this application; Figure 5 A schematic diagram of the specific functional units of the multimodal content generation module provided in an embodiment of this application; Figure 6 A schematic diagram of the specific functional units of the compliance review and content optimization module provided in an embodiment of this application; Figure 7 A schematic diagram of the specific functional units of the multi-platform adaptation and format conversion module provided in an embodiment of this application; Figure 8 A schematic diagram of the specific functional units of the content effect feedback module provided in an embodiment of this application; Figure 9 A schematic diagram illustrating the specific execution process of generating marketing content using an AI-driven marketing content automatic generation system provided in an embodiment of this application; Figure 10 A flowchart illustrating an AI-driven automatic marketing content generation method provided in an embodiment of this application; Figure 11 This is a schematic diagram of the structure of a computer device provided in an embodiment of this application. Detailed Implementation

[0011] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0012] To make the above-mentioned objectives, features and advantages of this application more apparent and understandable, the application will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0013] In one exemplary embodiment, such as Figure 1 As shown, an AI-driven automatic marketing content generation system is provided, including: The marketing data acquisition module M1 is used to acquire multi-source marketing analysis data of the target market (products / services to be marketed).

[0014] In another exemplary embodiment of this application, the marketing data collection module M1 is used to collect multi-source marketing analysis data, including internal enterprise marketing data, industry market data, user consumption behavior data, dissemination platform rule data, and competitor marketing data. After data cleaning and normalization, a standardized dataset is formed and stored in the marketing data resource library. After system initialization, it enters a standby state, providing a unified data source for subsequent operations. Specifically, as shown... Figure 2 As shown, the marketing data acquisition module M1 includes: a multi-source data acquisition unit M11, a data cleaning unit M12, a data normalization unit M13, and a data storage unit M14.

[0015] The M11 multi-source data acquisition unit is used to collect five major categories of data: internal enterprise marketing data, industry market data, user consumption behavior data, communication platform rule data, and competitor marketing data. Internal enterprise marketing data includes: product / service characteristics of the target market, historical marketing content, and historical conversion data; product characteristics include: product functions, parameters, materials, styles, efficacy, specifications, and packaging; service characteristics include: service processes, service duration, service scope, after-sales service, and value-added services; industry market data includes: industry trends, holiday hotspots, and marketing opportunities; this data is used to analyze marketing trends; user consumption behavior data includes user browsing, clicks, favorites, and purchase behavior; this data is used for user segmentation and demand analysis; communication platform rule data includes: content publishing guidelines, length limits, format requirements, and prohibited language guidelines for each communication platform; this data is used for compliance review and multi-platform format conversion; competitor marketing data includes: competitors' core marketing themes, content formats, and communication strategies; this data is used to analyze marketing trends.

[0016] The data cleaning unit M12 is used to perform missing value imputation, outlier removal, and duplicate value removal on the collected multi-source marketing analysis data. Mean imputation is used to impute numerical missing data, and mode imputation is used to impute categorical missing data. This is achieved through 3... σ The principle is to identify and remove numerical outliers.

[0017] The data normalization unit M13 is used to standardize cleaned data of different dimensions and scales, map numerical data to the [0, 1] interval, and perform one-hot encoding conversion on categorical data to form a standardized dataset with a unified format.

[0018] Data storage unit M14 is used to classify and store standardized datasets, build a marketing data resource library, and support on-demand retrieval and real-time updates for subsequent operations.

[0019] The User Profile and Demand Mining Module M2 is used to segment user groups based on user consumption behavior data and basic user information from multi-source marketing analysis data, using AI clustering algorithms. Based on the consumption behavior characteristics of each user group, it uses AI association rule mining algorithms to mine the user group's preferences for marketing content, and determines the marketing theme direction for the target audience based on the user group's preferences for marketing content.

[0020] In another exemplary embodiment of this application, the user profiling and demand mining module M2 is used to retrieve user consumption behavior data and basic user information (including user age, gender, region, occupation, consumption level, user account registration information, etc.) from the marketing data resource library, divide user groups using AI clustering algorithms, extract characteristics of each user group, mine user preferences for marketing content, and calculate the matching degree between the preset marketing direction and user preferences using a user preference matching degree formula to generate a "User Preference Demand Mining Report". Specifically, as shown... Figure 3 As shown, the user profiling and demand mining module M2 includes: user group clustering unit M21, user feature extraction unit M22, demand mining unit M23, and demand matching degree calculation unit M24.

[0021] The user group clustering unit M21 is used to segment users based on user consumption behavior data and basic user information from multi-source marketing analysis data, applying AI clustering algorithms. Specifically, it uses the K-means clustering algorithm to divide users into multiple distinct user groups, determine the core feature labels of each user group, and achieve the segmentation of different user groups.

[0022] The user feature extraction unit M22 is used to extract the consumption behavior characteristics of each user group. Specifically, the user feature extraction unit M22 is used to extract the core consumption behavior characteristics of each user group from the clustering results. Consumption behavior characteristics include spending power, consumption preferences, content browsing preferences, usage habits of communication platforms, and forms of accepting marketing content, etc., and feature vectors for each user group are constructed based on the consumption behavior characteristics of each user group.

[0023] The demand mining unit M23 is used to mine the preference needs of each user group for marketing content based on the consumption behavior characteristics of each user group and the basic characteristics of the target audience (product / service characteristics in the enterprise's internal marketing data). Specifically, the demand mining unit M23 uses AI association rule mining algorithms (such as the FP-Growth algorithm) to mine the core preference needs of each user group for marketing content, based on the characteristics of the product / service to be marketed and the characteristics of the user group (consumption behavior characteristics). Core preference needs include: content theme preference, content format preference, expression style preference, and release time preference, etc.

[0024] The demand matching degree calculation unit M24 is used to calculate the matching degree between the user group's preference needs for marketing content and multiple preset marketing directions. Based on the matching degree corresponding to the multiple preset marketing directions, the marketing theme direction corresponding to each user group is determined. Specifically, the demand matching degree calculation unit M24 is used to construct the user preference demand matching degree formula and calculate the matching degree between the preset marketing directions and the preference needs of each user group, which serves as the core basis for marketing theme positioning. The user preference demand matching degree formula is as follows: in, The value ranges from [0,1] to the user preference and demand matching degree. The larger the value, the higher the matching degree between the preset marketing direction and the user group's preference and demand. The number of dimensions for the preference and demand characteristics of marketing content, covering core dimensions such as content theme, content format, expression style, and release time; For the first The weights of each preference demand characteristic dimension are determined jointly by industry experience and historical data, satisfying... ; To pre-determine the marketing direction in the first The feature values ​​of each dimension range from [0, 1]. For the target user group in the first The feature values ​​of each dimension of preference demand, with a value range of [0, 1].

[0025] The Marketing Theme and Content Positioning Module M3 is used to determine the target marketing theme of the target audience based on the marketing theme direction and the marketing trend analysis results obtained from multi-source marketing analysis data, and to determine the marketing content positioning plan based on the target marketing theme and marketing characteristics of the target audience.

[0026] In another exemplary embodiment of this application, the marketing theme and content positioning module M3 is used to generate a candidate set of marketing themes based on the "User Needs Mining Report" and the results of marketing trend analysis. The final marketing theme (target marketing theme) is then determined through an AI-weighted scoring algorithm. Finally, the marketing content is positioned by combining the marketing characteristics (core selling points) of the product / service to be marketed. These marketing characteristics originate from internal enterprise marketing data, forming a "Marketing Content Positioning Scheme." Specifically, as shown... Figure 4 As shown, the marketing theme and content positioning module M3 includes: market trend analysis unit M31, theme candidate set generation unit M32, theme screening and determination unit M33, and content positioning unit M34.

[0027] The Market Trend Analysis Unit M31 is used to analyze multi-source marketing analysis data and identify the current marketing trend characteristics of the target product or service (i.e., the marketing trend analysis results). Specifically, the Market Trend Analysis Unit M31 analyzes industry market data and competitor marketing data to identify the current industry marketing trends, hot topics, and competitor marketing gaps of the target product or service, forming marketing trend characteristic tags for the target product or service.

[0028] The theme candidate set generation unit M32 is used to generate multiple marketing theme candidate schemes for each user group based on the marketing theme direction corresponding to each user group and the current marketing trend characteristics, forming a marketing theme scheme candidate set; each marketing theme candidate scheme includes a core theme, alternative themes, and core expression points of the theme.

[0029] The theme selection and determination unit M33 is used to score each marketing theme candidate scheme using AI weighted scoring algorithm based on user preference and demand matching degree, market fit, and competitor differentiation degree as screening indicators. Based on the scoring results (selecting the scheme with the highest score), the target marketing theme for each user group of the target audience is determined.

[0030] The content positioning unit M34 is used to determine the tone, key points, and collaborative direction of multimodal content for marketing content based on the target marketing theme and marketing characteristics of each user group. This forms a marketing content positioning plan for each user group, which determines the core theme, tone, and key points of the marketing content, serving as the core basis for generating marketing content.

[0031] The multimodal content generation module M4 is used to generate multimodal marketing content that meets the consistency verification requirements based on the marketing content positioning plan and through AI natural language processing algorithms. The consistency verification is used to check whether the key points, tone, and style of the multimodal marketing content are consistent.

[0032] In another exemplary embodiment of this application, the multimodal content generation module M4 is used to design the overall architecture of multimodal content representation according to the marketing content positioning scheme, generate initial multimodal marketing content such as text, image-text captions, and short social media posts through AI natural language processing algorithms, and perform collaborative verification to ensure that the tone and core information of each form of marketing content are consistent. Specifically, such as Figure 5 As shown, the multimodal content generation module M4 includes: The content architecture design unit M41 is used to design the overall content representation architecture of multimodal marketing content based on the marketing content positioning plan for each user group. This is to determine the core framework, length ratio, and content connection points of various forms of content, such as text copy, image and text captions, and short social media posts, to ensure that the tone of each form of content is consistent and the core information is consistent.

[0033] The multimodal content generation unit M42 is used to generate initial multimodal marketing content based on the overall content representation architecture and the key points of the marketing content, applying AI natural language processing algorithms. The text copy is divided into three parts: title, body, and conclusion; the image-text caption is divided into image description and core marketing text; and the short social media post is divided into three parts: concise title, core information, and interactive guidance.

[0034] The content collaboration verification unit M43 is used to perform collaboration verification on the initial multimodal marketing content to check whether the core expression points, communication tone, and expression style of the initial multimodal marketing content are consistent. The initial marketing content that fails the collaboration verification is adjusted to generate multimodal marketing content that meets the collaboration verification, ensuring that the multimodal marketing content forms a collaborative communication effect.

[0035] The compliance review and content optimization module M5 is used to review the compliance of multimodal marketing content on the dissemination platform, modify multimodal marketing content that fails the compliance review, apply AI marketing content quality assessment algorithms to conduct quality assessments in preset dimensions for multimodal marketing content that passes the compliance review, and optimize and adjust multimodal marketing content whose quality assessment results are lower than the preset quality score threshold.

[0036] In another exemplary embodiment of this application, the compliance review and content optimization module M5 is used to review the compliance of multimodal marketing content with the dissemination platform rules data in the multimodal marketing analysis data, mark and modify non-compliant content, score compliant content using an AI marketing content quality assessment algorithm, and perform targeted optimization on content whose scores do not reach the threshold until the quality meets the standard. Specifically, as shown... Figure 6 As shown, the compliance audit and content optimization module M5 includes: The compliance review unit M51 is used to build a compliance review rule base based on the communication platform rules data and industry compliance requirements of various communication platforms. Based on the compliance review rule base, it conducts a sentence-by-sentence and word-by-word compliance review of multimodal marketing content. Multimodal marketing content that fails the compliance review is marked and modified until the compliance review is passed. The compliance review rule base includes rules on prohibited words, rules on sensitive information, and rules on format specifications.

[0037] The content quality assessment unit M52 is used to evaluate the quality of multimodal marketing content that has passed compliance review using an AI marketing content quality assessment algorithm across preset dimensions, deriving quality scores for each dimension. Specifically, the content quality assessment unit M52 is used to construct an AI marketing content quality assessment formula to evaluate the quality of content that has passed compliance review from five dimensions: content completeness, information accuracy, fluency of expression, prominence of selling points, and user appeal. The AI ​​marketing content quality assessment formula is as follows: in, The quality score for marketing content is determined by a value ranging from [0, 10]. A higher value indicates higher content quality. These are the serial numbers for the quality assessment dimensions, covering content completeness, information accuracy, fluency of expression, prominence of selling points, and user appeal; For the first The weights of each evaluation dimension satisfy the following: ; For the first The scores for each evaluation dimension, ranging from [0, 10], are determined by a combination of preset scoring criteria and human-assisted scoring.

[0038] The targeted optimization unit M53 is used to perform targeted optimization adjustments on multimodal marketing content based on preset dimensions where the quality assessment score is lower than a preset quality score threshold, until the quality assessment score is higher than the corresponding preset quality score threshold, thus obtaining the target multimodal marketing content. Specifically, the targeted optimization unit M53 is used to set the quality score threshold; when the marketing content quality assessment score... When the score is below the preset quality score threshold, targeted optimization is carried out based on the scores of each dimension. The content of the dimensions with low scores is supplemented, modified or rewritten. After optimization, the quality assessment is carried out again until the quality assessment score reaches the corresponding threshold requirement.

[0039] In this application, the design of the compliance review and content optimization module M5 addresses the problem that in the current traditional marketing content generation model, content quality assessment and compliance review rely entirely on manual processes, resulting in inconsistent review standards, easy inclusion of illegal information and omissions in expression, and increased risks associated with content publication.

[0040] In another exemplary embodiment of this application, the AI-driven marketing content automatic generation system further includes: a multi-platform adaptation and format conversion module M6, used to extract the core content publishing characteristics of each dissemination platform based on the dissemination platform rule data in multi-source marketing analysis data, and to construct a content publishing adaptation rule library in combination with the dissemination platform rule data. Based on the content publishing adaptation rule library, the system performs multi-platform format conversion (format, length, and expression form adaptation conversion) on marketing content that meets the quality assessment standards, evaluates the adaptation effect through an AI content platform adaptation algorithm, and re-optimizes content that does not meet the adaptation standards to generate exclusive marketing content for each platform. Specifically, as shown below... Figure 7 As shown, the M6 ​​multi-platform adaptation and format conversion module includes: The platform feature extraction unit M61 is used to extract the content publishing features of each communication platform based on the communication platform rule data in the multi-source marketing analysis data. The content publishing features include content length restrictions, format requirements, expression style preferences, and publishing format requirements.

[0041] The adaptation rule building unit M62 is used to build a content publishing adaptation rule library based on the content publishing characteristics and rule data of each dissemination platform. The content publishing adaptation rule library includes rules for adjusting the length, format conversion, expression style adaptation, and publishing format adaptation.

[0042] The format conversion unit M63 is used to perform platform format conversion on target multimodal marketing content based on the content publishing adaptation rule library (including adding or deleting text, adjusting format, adapting expression style, and converting publishing format), generating exclusive multimodal marketing content for each communication platform.

[0043] The platform adaptability calculation unit M64 is used to calculate the content platform adaptability of customized multimodal marketing content based on the content publishing characteristics of each dissemination platform, using an AI content platform adaptability algorithm. The platform adaptability calculation unit M64 is also used to construct the content platform adaptability evaluation formula corresponding to the AI ​​content platform adaptability algorithm. The specific formula is as follows: in, This represents the content platform compatibility score, with a value range of [0, 1]. A higher value indicates a higher compatibility between the marketing content and the communication platform. The number of dimensions representing the characteristics of the dissemination platform, covering core dimensions such as length, format, style, and publishing method; For the first The propagation platform feature values ​​are defined in each dimension, with values ​​ranging from [0, 1]. For the converted marketing content in the first Each dimension contains feature values, ranging from [0, 1]. When the content platform adaptability... When the content's adaptability falls below a preset threshold, the format conversion and adaptation optimization are performed again. The format conversion unit M63 is also used to re-convert the platform format of target multimodal marketing content whose content platform adaptability is below a preset threshold.

[0044] In this application, the design of the multi-platform adaptation and format conversion module M6 solves the problem that in the current traditional marketing content generation model, in multi-platform publishing scenarios, manual adjustments to the format, length, and expression of the same marketing content are required, resulting in low adaptation efficiency and easy content tone deviation. This application can achieve automated cross-platform adaptation and tone consistency.

[0045] In another exemplary embodiment of this application, the AI-driven automatic marketing content generation system further includes: a content effect feedback module M7, used to collect real-time data on the dissemination, conversion, and user feedback of marketing content after it is published on the corresponding dissemination platform, conduct a comprehensive effect evaluation, and adjust the core parameters of each module and iteratively optimize the relevant AI algorithm models applied based on the evaluation results, forming a closed-loop management of marketing content generation. Specifically, as shown in the example... Figure 8 As shown, the content effect feedback module M7 includes: The M71 performance data collection unit is used to collect end-to-end performance data after the release of exclusive multimodal marketing content across various communication platforms. This end-to-end performance data includes dissemination data, conversion data, and user feedback data. Dissemination data includes impressions, clicks, shares, and comments; conversion data includes saves, inquiries, purchases, and repeat purchases; and user feedback data includes comment sentiment and user suggestions.

[0046] The effectiveness evaluation system construction unit M72 is used to comprehensively score the content feedback effect of each dissemination platform based on the full-link effectiveness data of each dissemination platform, using dissemination efficiency, conversion efficiency, and user satisfaction as evaluation indicators, and to obtain the content feedback effect score. It can identify the content characteristics with good content feedback effect and the problem points with poor effect.

[0047] In another exemplary embodiment of this application, the content effect feedback module M7 further includes: a module parameter optimization unit M73 and an algorithm model iteration unit M74.

[0048] Among them, the module parameter optimization unit M73 is used to adjust the core parameters of the aforementioned modules in reverse based on the content feedback effect score (such as the weights in the user preference demand matching formula). Weights in the content quality assessment formula (e.g., the dimensions and relevant thresholds for platform feature extraction) make the output results of each module more closely match actual marketing needs.

[0049] The algorithm model iteration unit M74 is used to iteratively optimize the lightweight AI and statistical algorithms used in various modules of the system. These lightweight AI and statistical algorithms include feature dimensions for user group clustering, confidence thresholds for association rule mining, and core parameters for natural language processing. By combining content feedback data, the accuracy and adaptability of the relevant AI algorithms are continuously optimized, thereby achieving a continuous improvement in the marketing content generation capability.

[0050] In this application, the design of the content effect feedback module M7 solves the problem that the current traditional marketing content generation model lacks an effective content effect feedback and iteration mechanism, and the dissemination data and conversion data after the marketing content is released cannot guide subsequent content creation, thus forming data silos.

[0051] In this application, the modular design of the AI-driven marketing content automatic generation system solves the problems of the lack of standardization in the content generation process, the disconnect between various links from demand analysis to content implementation, the low overall generation efficiency, and the high labor costs in the current traditional marketing content generation model.

[0052] This application utilizes lightweight AI algorithms combined with multi-source marketing analysis data mining to automate the entire marketing content process, from data collection, demand mining, theme positioning, content generation, optimization and review, platform adaptation to performance feedback. Simultaneously, standardized module design and algorithm models ensure the personalization, compliance, and multi-platform compatibility of marketing content, balancing practicality and feasibility. The AI-driven component is kept at a low level to avoid problems such as high implementation costs and poor adaptability caused by over-reliance on complex models, thus meeting the actual marketing needs of small and medium-sized enterprises.

[0053] This application abandons the traditional experience-driven creation model, utilizing multi-source marketing analysis data to mine user preferences and marketing trends, ensuring the personalization and market relevance of marketing content and effectively avoiding content homogenization. It achieves standardized control from data collection, demand mining, theme positioning, content generation, optimization review, platform adaptation to effect feedback, with the output results of each module serving as the basis for subsequent modules, significantly improving the overall efficiency of marketing content generation. It ensures the consistency of core information and communication tone across various content formats, including text, graphic captions, and short social media posts, enabling multimodal marketing content to achieve synergistic brand communication effects. It establishes unified review and scoring standards, effectively reducing the risk of unauthorized publication of marketing content, while targeted optimization improves the overall quality of content, ensuring information accuracy and user appeal. It eliminates the need for repetitive format adjustments to the same marketing content, enabling rapid cross-platform deployment of marketing content, improving cross-platform dissemination efficiency, and reducing labor costs. Furthermore, it uses marketing content dissemination and conversion data to optimize the parameters and algorithms of each module, allowing the system's content generation capabilities to continuously improve with market changes and user needs, forming a sustainable marketing content generation capability.

[0054] like Figure 9 As shown below, taking a holiday marketing scenario applied to e-commerce beauty brands, which requires generating multimodal marketing content adapted to e-commerce platforms, social media platforms, and short video platforms, as an example, the execution flow of each module of the AI-driven automatic marketing content generation system in this application is illustrated: Step (a1): After system initialization, the marketing data collection module M1 is used to collect multi-source marketing analysis data of the beauty brand, including product characteristics (such as moisturizing and anti-aging effects, lotion texture, 50ml size product characteristics), historical holiday marketing data, e-commerce platform user consumption behavior data, holiday marketing rules of various platforms, and competitor holiday marketing content. After cleaning and normalization, a standardized dataset is formed and stored in the marketing data resource library.

[0055] Step (a2): Utilize the user profiling and demand mining module M2 to retrieve user consumption behavior data. Combine this with basic user information (e.g., students aged 18-25, working women aged 25-40, and beauty enthusiasts aged 18-35). Using the K-means clustering algorithm, users are divided into three main categories: young students, working women, and beauty enthusiasts. Extract consumption preferences, content browsing preferences, and platform usage habits for each user group to uncover their core needs (preference needs) for holiday marketing content. For example, young students prefer cost-effective themes and short copywriting, working women prefer efficacy themes and graphic formats, and beauty enthusiasts prefer ingredient themes and detailed text formats. Calculate the matching degree between the two preset marketing directions, "Holiday Limited Edition and High Cost-Effectiveness" and "Holiday Limited Edition and Core Efficacy," and the preference needs of each user group using the user preference need matching formula. Determine "Holiday Limited Edition and Core Efficacy" as the core marketing direction and generate the "User Demand Mining Report."

[0056] Step (a3): Using the Marketing Theme and Content Positioning Module M3, based on the "User Needs Mining Report" and the beauty industry's holiday marketing trends, generate candidate marketing theme schemes such as "Beautiful Skin Renewal, Holiday Selection" and "Holiday Limited Edition, Efficacy Skin Renewal". After weighted scoring and screening, "Holiday Limited Edition, Efficacy Skin Renewal" is selected as the final marketing theme. Combining the brand's core selling points of moisturizing and anti-aging (derived from the company's internal marketing data), the content communication tone is determined to be high-end and sophisticated, and the core expression points are holiday limited edition, core efficacy, and holiday discounts, thus forming the "Marketing Content Positioning Scheme".

[0057] Step (a4): Using the multimodal content generation module M4, design a multimodal content representation architecture based on the "Marketing Content Positioning Scheme": text copy for e-commerce platforms includes "title + product introduction + efficacy description + holiday discount + closing guidance", text and image captions for social media platforms include "image description + core efficacy + holiday benefits", and short social media posts for short video platforms include "concise title + core discount + interactive guidance". Based on the content representation architecture, generate initial multimodal content in various forms using AI natural language processing algorithms, and perform collaborative verification to ensure that all content revolves around "holiday limited edition, core efficacy" and maintains a consistent tone.

[0058] Step (a5): Using the compliance review and content optimization module M5, based on the dissemination platform rule data of e-commerce, social media, and short video platforms, the initial multimodal content is reviewed, non-standard terms are marked and modified, and the compliant content is scored using the marketing content quality assessment formula. For the issue of low scores in the "selling point prominence" dimension, the core efficacy parts of each form of content are supplemented and modified, and the quality score after optimization reaches the threshold requirement.

[0059] Step (a6): Utilize the multi-platform adaptation and format conversion module M6 to extract the core features of e-commerce platforms, social media platforms, and short video platforms (derived from the platform rules data). Based on the adaptation rule library, perform format conversion on content that meets the quality standards: e-commerce platforms retain complete text copy and accompanying images, social media platforms streamline the text and highlight promotional information, and short video platforms convert the content into conversational short dynamics. Through the content platform adaptation formula evaluation, the adaptation degree of each communication platform reaches the preset adaptation degree threshold, generating exclusive marketing content for each communication platform.

[0060] Step (a7): Publish the exclusive marketing content for each communication platform to the corresponding platform. Use the performance data collection unit M71 to collect real-time data such as exposure, clicks, purchases, and comment sentiment. Use the performance evaluation system construction unit M72 to conduct a comprehensive evaluation of the content feedback effect. It was found that the conversion efficiency of the short video platform content was relatively high, while the user feedback on the selling points of the e-commerce platform content was not clear enough. Based on this result, the system reversely adjusted the weight of the "content format" dimension in the user preference demand matching formula and the weight of the "information accuracy" dimension in the content quality evaluation formula. At the same time, iteratively optimized the text expression algorithm of AI multimodal content generation, making the selling points of the subsequently generated e-commerce platform content clearer and further improving the conversion ability of the short video platform content.

[0061] like Figure 9 As shown below, taking the opening marketing scenario of offline catering stores as an example, which requires generating multimodal marketing content adapted to local life platforms, community social platforms, and offline store posters, the execution flow of each module of the AI-driven automatic marketing content generation system in this application is explained: Step (b1): After system initialization, the marketing data collection module M1 is used to collect multi-source marketing analysis data for the restaurant, including service characteristics (such as dine-in, takeout, and post-meal packaging), product characteristics (local home-style dishes, specialty snacks, dishes with an average price of 50 yuan per person, and average order value), local catering industry data, surrounding user consumption behavior data, local life platform rules, and marketing data of surrounding competitors' openings. After cleaning and normalization, a standardized dataset is formed and stored in the marketing data resource library.

[0062] Step (b2): Utilize the user profiling and demand mining module M2 to retrieve surrounding user consumption behavior data. Combine this with basic user information (residents of surrounding communities, office workers, and university students) to categorize users into three main groups: surrounding residents, office workers, and students. Uncover the core needs of each group: surrounding residents prefer cost-effectiveness and home-style cooking themes; office workers prefer fast food and convenience themes; and students prefer unique and discounted themes. Calculate the matching degree of each preset marketing direction using the user preference demand matching formula, determine "Grand Opening Special Offers, Local Flavors" as the core marketing direction, and generate the "User Demand Mining Report".

[0063] Step (b3): ​​Using the Marketing Theme and Content Positioning Module M3, based on user preferences and local restaurant opening marketing trends, determine "Grand Opening Specials, Authentic Local Flavors" as the final marketing theme. Combine the store's local specialty dishes and opening discount core selling points (derived from internal marketing data), determine the content communication tone to be approachable and affordable, with the core message being local specialties, opening discounts, and in-store benefits, thus forming the "Marketing Content Positioning Plan".

[0064] Step (b4): Using the multimodal content generation module M4, design a multimodal content representation architecture based on the "Marketing Content Positioning Scheme": text copy for local life platforms includes "title + store introduction + dish features + opening discount + in-store guidance", short copy for community social platforms includes "title + core offers + store address", and text for offline posters includes "large font title + discount information + introduction of featured dishes"; generate initial multimodal marketing content in various forms through AI natural language processing algorithms and perform collaborative verification to ensure consistency of core information and a friendly tone.

[0065] Step (b5): Using the compliance review and content optimization module M5, the initial multimodal marketing content is reviewed for compliance based on the rules and data of the dissemination platform. Illegal content is removed, and the content quality is scored using the marketing content quality assessment formula. For the issue of low scores in the "user appeal" dimension, the preferential information section of each form of content is significantly modified. After content optimization, the quality score meets the standard.

[0066] Step (b6): Utilize the M6 ​​multi-platform adaptation and format conversion module to extract the core features (derived from the dissemination platform rule data) of local life platforms, community social platforms, and offline posters, and perform format conversion on the multimodal marketing content: retain complete information for local life platforms, simplify the community social platforms into short, life-oriented copy, and simplify the text of offline posters into large headlines + core discounts; evaluate through the content platform adaptation formula, and if the adaptation of each dissemination platform meets the standard, generate exclusive marketing content for each platform / scenario.

[0067] Step (b7): Publish exclusive marketing content to the corresponding communication platforms and create offline posters. Collect feedback data such as exposure, store visits, spending, and user reviews. After comprehensive evaluation, it was found that the content on the community social platform had the highest conversion rate to stores, and the user feedback on the introduction of featured dishes on the offline posters was not clear enough. Based on this result, the system adjusted the core parameters of each module and iteratively optimized the architecture design algorithm of AI content generation, so that the text of the subsequently generated offline posters would highlight the featured dishes, and the discount information on the community social platform would be more in line with the needs of users around the store.

[0068] Based on the same inventive concept, this application also provides an AI-driven marketing content automatic generation method implemented based on the aforementioned AI-driven marketing content automatic generation system. The solution provided by this method is similar to the implementation described in the system above. Therefore, the specific limitations in the embodiments of the AI-driven marketing content automatic generation method provided below can be found in the limitations of the AI-driven marketing content automatic generation system described above, and will not be repeated here.

[0069] In one exemplary embodiment, an AI-driven method for automatically generating marketing content is provided, such as... Figure 10 As shown, it includes: S1: Obtain multi-source marketing analysis data for the target audience; multi-source data includes internal enterprise marketing data, industry market data, user consumption behavior data, communication platform rule data, and competitor marketing data.

[0070] S2: Based on user consumption behavior data and basic user information (such as age, gender, region, occupation, and consumption level) from multi-source marketing analysis data, user groups are segmented. Based on the consumption behavior characteristics of each user group (including spending power, consumption preferences, content browsing preferences, communication platform usage habits, and marketing content acceptance formats), AI association rule mining algorithms are applied to mine the preference needs of each user group for marketing content. Based on the preference needs of each user group for marketing content, the marketing theme direction for the target audience is determined.

[0071] S3: Determine the target marketing theme of the target audience based on the marketing theme direction and the marketing trend analysis results obtained from multi-source marketing analysis data. Determine the marketing content positioning plan based on the target marketing theme of the target audience and the marketing characteristics in the company's internal marketing data.

[0072] S4: Based on the marketing content positioning plan, use AI natural language processing algorithms to generate multimodal marketing content that meets the consistency verification; the consistency verification is used to check whether the key points, tone, and style of the multimodal marketing content are consistent.

[0073] S5: Based on the dissemination platform rules data in the multimodal marketing analysis data, conduct a dissemination platform compliance review of the multimodal marketing content. Modify the multimodal marketing content that fails the compliance review. For the multimodal marketing content that passes the compliance review, apply the AI ​​marketing content quality assessment algorithm to conduct a quality assessment in preset dimensions. Optimize and adjust the multimodal marketing content whose quality assessment results are lower than the preset quality score threshold.

[0074] In one exemplary embodiment, a computer device is provided, which may be a server or a terminal, and its internal structure diagram may be as follows. Figure 11 As shown, the computer device includes a processor, memory, input / output interfaces (I / O), and a communication interface. The processor, memory, and I / O interfaces are connected via a system bus, and the communication interface is also connected to the system bus via the I / O interfaces. The processor provides computational and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system, computer programs, and a database. The internal memory provides the environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The database stores relevant data from the AI-driven automatic marketing content generation process. The I / O interfaces are used for information exchange between the processor and external devices. The communication interface is used for communication with external terminals via a network connection. When the computer program is executed by the processor, it implements an AI-driven automatic marketing content generation method.

[0075] Figure 11 The structures shown are merely block diagrams of some structures related to the present application and do not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than shown in the figures, or combine certain components, or have different component arrangements. In an exemplary embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the steps in the above-described method embodiments.

[0076] In one exemplary embodiment, a computer-readable storage medium is provided storing a computer program that, when executed by a processor, implements the steps described in the method embodiments above.

[0077] In one exemplary embodiment, a computer program product is provided, including a computer program that, when executed by a processor, implements the steps in the method embodiments described above.

[0078] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of related data must comply with relevant regulations and be authorized by the owner of the corresponding device.

[0079] Those skilled in the art will understand that all or part of the processes in the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments described above. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM).

[0080] The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0081] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0082] This document uses specific examples to illustrate the principles and implementation methods of this application. The descriptions of the above embodiments are only for the purpose of helping to understand the methods and core ideas of this application. Furthermore, those skilled in the art will recognize that, based on the ideas of this application, there will be changes in the specific implementation methods and application scope. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. An AI-driven automatic marketing content generation system, characterized in that, include: The marketing data acquisition module is used to acquire multi-source marketing analysis data of the target audience; The user profiling and demand mining module is used to segment user groups based on user consumption behavior data and basic user information from multi-source marketing analysis data, using AI clustering algorithms. Based on the consumption behavior characteristics of each user group, AI association rule mining algorithms are used to mine the user group's preferences for marketing content. Based on the user group's preferences for marketing content, the marketing theme direction for the target audience is determined. The marketing theme and content positioning module is used to determine the target marketing theme of the target marketing object based on the marketing theme direction of the target marketing object and the marketing trend analysis results obtained based on multi-source marketing analysis data, and to determine the marketing content positioning plan based on the target marketing theme and marketing characteristics of the target marketing object. The multimodal content generation module is used to generate multimodal marketing content that meets the consistency verification requirements by applying AI natural language processing algorithms based on the marketing content positioning plan. The consistency verification is used to check whether the key points, tone, and style of the multimodal marketing content are consistent. The compliance review and content optimization module is used to review the compliance of multimodal marketing content with the dissemination platform rules data in the multimodal marketing analysis data. It modifies multimodal marketing content that fails the compliance review, and applies an AI marketing content quality assessment algorithm to conduct a quality assessment of preset dimensions for multimodal marketing content that passes the compliance review. It also optimizes and adjusts multimodal marketing content whose quality assessment results are lower than the preset quality score threshold.

2. The AI-driven automatic marketing content generation system according to claim 1, characterized in that, The user profiling and demand mining module includes: The user group clustering unit is used to divide user groups based on user consumption behavior data and basic user information from multi-source marketing analysis data, using AI clustering algorithms. The user feature extraction unit is used to extract the consumption behavior characteristics of each user group. The demand mining unit is used to mine the preference needs of each user group for marketing content by combining the consumption behavior characteristics of each user group with the basic characteristics of the target audience and applying AI association rule mining algorithms. The demand matching degree calculation unit is used to calculate the matching degree between the user group's preference demand for marketing content and multiple preset marketing directions, and determine the marketing theme direction corresponding to each user group based on the matching degree corresponding to multiple preset marketing directions.

3. The AI-driven automatic marketing content generation system according to claim 1, characterized in that, The marketing theme and content positioning module includes: The market trend analysis unit is used to analyze multi-source marketing analysis data to obtain the current marketing trend analysis results for the target market. The theme candidate set generation unit is used to generate multiple marketing theme candidate schemes for each user group based on the marketing theme direction corresponding to each user group and the current marketing trend analysis results. The theme selection and determination unit uses user preference and demand matching, market fit, and competitor differentiation as selection indicators. It uses AI weighted scoring method to score each marketing theme candidate plan and determines the target marketing theme for each user group of the target audience based on the scoring results. The content positioning unit is used to determine the tone, key points, and collaborative direction of multimodal content for marketing content based on the target marketing theme and marketing characteristics of each user group, thus forming a marketing content positioning plan for each user group.

4. The AI-driven automatic marketing content generation system according to claim 1, characterized in that, The multimodal content generation module includes: The content architecture design unit is used to design the overall content representation architecture of multimodal marketing content based on the marketing content positioning scheme of each user group; The multimodal content generation unit is used to generate initial multimodal marketing content by applying AI natural language processing algorithms based on the overall content representation architecture and the key points of the marketing content. The content collaboration verification unit is used to perform collaboration verification on the initial multimodal marketing content, adjust the initial marketing content that fails the collaboration verification, and generate multimodal marketing content that meets the collaboration verification.

5. The AI-driven automatic marketing content generation system according to claim 1, characterized in that, The compliance review and content optimization module includes: The compliance review unit is used to build a compliance review rule library based on the communication platform rules data and industry compliance requirements of each communication platform, conduct communication platform compliance review of multimodal marketing content based on the compliance review rule library, and modify multimodal marketing content that fails the compliance review until the compliance review is passed; The content quality assessment unit is used to evaluate the quality of multimodal marketing content that has passed compliance review using AI marketing content quality assessment algorithms across preset dimensions, and to obtain quality assessment scores for each preset dimension. The targeted optimization unit is used to optimize and adjust multimodal marketing content based on preset dimensions where the quality assessment score is lower than the preset quality score threshold, until the quality assessment score is higher than the corresponding preset quality score threshold, thus obtaining the target multimodal marketing content.

6. The AI-driven automatic marketing content generation system according to claim 5, characterized in that, The AI-driven marketing content automatic generation system also includes: a multi-platform adaptation and format conversion module; the multi-platform adaptation and format conversion module includes: The platform feature extraction unit is used to extract the content publishing features of each communication platform based on the communication platform rule data in the multi-source marketing analysis data. The adaptation rule building unit is used to build a content publishing adaptation rule library based on the content publishing characteristics and rule data of each dissemination platform. The format conversion unit is used to convert the target multimodal marketing content into platform formats based on the content publishing adaptation rule library, generating exclusive multimodal marketing content for each dissemination platform; it is also used to re-convert the platform format of target multimodal marketing content whose content platform adaptation is lower than the preset adaptation threshold. The platform adaptability calculation unit is used to calculate the content platform adaptability of exclusive multimodal marketing content based on the content publishing characteristics of each communication platform and by applying an AI content platform adaptability algorithm.

7. The AI-driven automatic marketing content generation system according to claim 1, characterized in that, The AI-driven marketing content automatic generation system also includes: a content performance feedback module; the content performance feedback module includes: The performance data collection unit is used to collect full-link performance data after the release of exclusive multimodal marketing content on various communication platforms; The effectiveness evaluation system construction unit is used to comprehensively score the content feedback effect of each dissemination platform based on the full-link effectiveness data of each dissemination platform, using dissemination efficiency, conversion efficiency, and user satisfaction as evaluation indicators, and to obtain the content feedback effect score.

8. A method for automatically generating marketing content based on AI, characterized in that, include: Obtain multi-source marketing analysis data from the target audience; Based on user consumption behavior data and basic user information from multi-source marketing analysis data, AI clustering algorithms are applied to segment user groups. Based on the consumption behavior characteristics of each user group, AI association rule mining algorithms are applied to mine the user group's preference needs for marketing content. Based on the user group's preference needs for marketing content, the marketing theme direction for the target audience is determined. The target marketing theme of the target audience is determined based on the marketing theme direction and the marketing trend analysis results obtained from multi-source marketing analysis data. The marketing content positioning plan is determined based on the target marketing theme and marketing characteristics of the target audience. Based on the marketing content positioning plan, AI natural language processing algorithms are used to generate multimodal marketing content that meets the consistency verification. The consistency verification is used to check whether the key points, tone, and style of the multimodal marketing content are consistent. Based on the dissemination platform rules data in the multimodal marketing analysis data, the dissemination platform compliance review of multimodal marketing content is conducted. Multimodal marketing content that fails the compliance review is modified. For multimodal marketing content that passes the compliance review, an AI marketing content quality assessment algorithm is applied to conduct a quality assessment in preset dimensions. Multimodal marketing content whose quality assessment results are lower than the preset quality score threshold is optimized and adjusted.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by a processor, the computer program implements the AI-driven automatic marketing content generation method described in claim 8.

10. A computer program product, comprising a computer program, characterized in that, When executed by a processor, the computer program implements the AI-driven automatic marketing content generation method described in claim 8.

Images