Product marketing copy automatic generation method and system

Abstract

The application provides a commodity marketing copy automatic generation method and system, and belongs to the technical field of data processing under the internet industry. The method comprises the following steps: obtaining a structured selling point list generated by a product information and a competitive product information according to a selling point analysis node; for any subsequent execution node, the following operations are performed: obtaining product supplementary information input by a user at a current execution node and node output results of at least one executed node; generating node output results of the current execution node according to the product supplementary information and the node output results of the at least one executed node; selecting to jump to a next execution node, or when the current execution node is a copy generation node, determining that the node output results of the current execution node are reference output results; performing compliance verification on the reference output results according to a target market, and outputting a marketing copy corresponding to the reference output results. The structured generation of commodity marketing copies for different target markets can be realized, and the copy generation quality is improved.

Description

Technical Field

[0001] This application belongs to the field of data processing in the Internet industry, specifically involving a method and system for automatically generating product marketing copy. Background Technology

[0002] Currently, product marketing copy generation primarily relies on manual writing or template-based automated generation. The former is limited by inefficiency and high costs, making it difficult to scale, especially in the multilingual and multi-market cross-border e-commerce environment. While the latter can improve efficiency, it suffers from key issues such as inconsistent copy quality, difficulty in ensuring brand consistency, and insufficient integration of product selling points with marketing scenarios. Furthermore, existing methods lack the ability to adapt to the cultural preferences, language habits, and compliance requirements of target countries / regions, making it difficult to achieve controllable and precise copy generation based on structured constraints (such as region, style, and templates). This results in fragmented marketing content across different markets, hindering the efficiency and effectiveness of brand globalization operations. Summary of the Invention

[0003] This application provides a method and system for automatically generating product marketing copy, so as to realize the structured generation of product marketing copy for different target markets and improve the quality of copy generation.

[0004] This application provides a method for automatically generating product marketing copy, including:

[0005] Obtain product information of the target product input by the user at the selling point analysis node, as well as competitor information of at least one set of competitors of the target product; and generate a structured selling point list at the selling point analysis node based on the product information and the competitor information, wherein the structured selling point list includes the core selling points of the target product, supplementary explanations of the core selling points, and image content plans for the core selling points;

[0006] For any subsequent execution node, the following operations are performed: Obtain supplementary product information input by the user at the current execution node, and the node output results of at least one executed node, wherein the node output results of the at least one executed node include the structured selling point list generated by the selling point analysis node; generate the node output result of the current execution node based on the supplementary product information and the node output results of the at least one executed node; and determine the node output result of the current execution node as a reference output result, depending on the user's selection to jump to the next execution node, or when the current execution node is a copywriting generation node; each subsequent execution node sequentially includes a node for generating image tags, a node for generating a storyline for the target product, a style matching node for generating and determining the style attributes of the target product, a template matching node for determining the output template, and a copywriting generation node for generating marketing copy.

[0007] The reference output results are validated for compliance based on the target market to which the target product belongs, and the marketing copy corresponding to the reference output results is output after the validation is passed.

[0008] According to the product marketing copy automatic generation method provided in this application, before jumping to the next execution node based on the user's selection, the method further includes: determining whether the node output result of the current execution node has been obtained; if the node output result has not been obtained, determining whether the current execution node is a necessary execution node; if it is a necessary execution node, outputting a prompt message, the prompt message being used to indicate the necessity of executing the current execution node; after outputting the prompt message, if the node output result has not been obtained again, and the user's instruction to jump to the next execution node is received, generating the default output result of the current execution node based on the node output results of at least one already executed node, and jumping to the next execution node.

[0009] In one possible embodiment, generating the default output result of the currently executing node based on the node output results of the at least one executed node includes: determining whether the node output result of the currently executing node depends on the supplementary information; if it depends on the supplementary information, generating default supplementary information based on the product information and the competitor information, and generating the default output result based on the node output results of the at least one executed node and the default supplementary information.

[0010] In one possible embodiment, generating the target product storyline includes: obtaining supplementary product information input by the user at the storyline generation node, the supplementary product information including attribute information of the core selling points; generating a competitor storyline based on the attribute information and competitor image tags of at least one set of competitors output by the image tag generation node; and generating the target product storyline based on the attribute information, the structured selling point list, and the competitor storyline.

[0011] In one possible embodiment, generating a competitor storyline from multiple competitor images and competitor image tags for each competitor image, generated based on the attribute information and the image tags, includes: determining a first target scenario based on the attribute information; generating a multi-act story structure under the target scenario based on the multiple competitor images and the competitor image tags, wherein the multi-act story structure constitutes the competitor storyline, each act corresponding to a core selling point, the arguments and justifications for the core selling point, and the narrative techniques, image style, and emotional triggers of the core selling point.

[0012] In one possible embodiment, generating the target product storyline based on the attribute information, the structured selling point list, and the competitor storyline includes: determining the target product's advantages against the competitor for each core selling point based on the attribute information and the structured selling point list; and generating the target product storyline based on the advantages and the competitor storyline.

[0013] In one possible embodiment, the copy generation node performs the following operations: generates text information corresponding to each story structure based on the target product storyline; determines at least one target image corresponding to each story structure based on the text information and the image tags; and generates marketing copy information corresponding to each story structure based on the text information and the at least one target image.

[0014] In one possible embodiment, determining at least one target image corresponding to each scene story structure based on the text information and the image tags includes: performing the following operations for each scene story structure: selecting a first group of images from reference images whose image tags include scene image types; selecting a second group of images from the first group of images whose product selling points, as included in the image tags, correspond to the target context; selecting a third group of images from the second group of images whose image content summary, as included in the image tags, includes key image elements, the key image elements including image elements associated with the core selling points; performing a quality score on the third group of images, and determining the images whose quality score is higher than a preset value as target images.

[0015] In one possible embodiment, the quality scoring of the third set of images includes: performing the following operations for each image in the third set of images: determining the image sharpness of the image and the relevance of the image to the core selling point; determining an image quality score based on the image sharpness and the relevance to the selling point; determining image risk based on the target market and the marketing copy placement platform; performing a safety quality score on the third set of images based on the image risk; and determining a final quality score for the image based on the image quality score and the safety quality score.

[0016] This application also provides an automatic product marketing copy generation system, including:

[0017] The selling point analysis module is used to obtain product information of the target product input by the user at the selling point analysis node, as well as the competitor information of at least one set of competitors of the target product; and to generate a structured selling point list based on the product information and the competitor information. The structured selling point list includes the core selling points of the target product, supplementary explanations of the core selling points, and image content plans for the core selling points.

[0018] Each subsequent execution module is used to perform the following operations: obtain supplementary product information input by the user at the current execution node, and the node output results of at least one executed module, wherein the node output results of the at least one executed module include a structured list of selling points generated by the selling point analysis module; generate the node output result of the current execution module based on the supplementary product information and the node output results of the at least one executed module; and determine the node output result of the current execution module as a reference output result, depending on the user's selection to jump to the next execution module, or when the current execution module is a copywriting generation module; each subsequent execution module sequentially includes an image tag generation module for outputting image tags, a storyline generation module for generating the target product storyline, a style matching module for generating and determining the style attributes of the target product, a template matching module for determining the output template, and the copywriting generation module for generating marketing copy;

[0019] The compliance verification module is used to perform compliance verification on the reference output result according to the target market to which the target product belongs, and output the marketing copy corresponding to the reference output result after the verification is passed.

[0020] This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein the processor executes the computer program to implement any of the above-described methods for automatically generating product marketing copy.

[0021] This application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements any of the above-described methods for automatically generating product marketing copy.

[0022] This application also provides a computer program product, including a computer program that, when executed by a processor, implements any of the above-described methods for automatically generating product marketing copy.

[0023] The product marketing copy automatic generation method and system provided in this application generates a structured list of selling points at the selling point analysis node, and then generates corresponding output results based on user selections at different execution nodes such as image tag generation, storyline generation, style matching, template matching, and copy generation. This achieves structured generation of product marketing copy for different target markets, improving the quality of the generated copy. The generated marketing copy not only effectively integrates with the core selling points of the product but also adapts to different target markets, improving the compliance of the copy. Attached Figure Description

[0024] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0025] Figure 1 This is a flowchart illustrating a method for automatically generating product marketing copy provided in this application.

[0026] Figure 2 This is one of the task creation diagrams provided in this application.

[0027] Figure 3 This is the second illustration of task creation provided in this application.

[0028] Figure 4 This is the third illustration of task creation provided in this application.

[0029] Figure 5 This is a schematic diagram of the task node skipping results provided in this application.

[0030] Figure 6 This is a schematic diagram of the competitor's storyline generation provided in this application.

[0031] Figure 7 This is a schematic diagram of the components of an automatic product marketing copy generation system provided in this application.

[0032] Figure 8 This is a schematic diagram of the structure of an electronic device provided in this application. Detailed Implementation

[0033] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0034] The terms "first," "second," etc., in the specification, claims, and accompanying drawings of this application are used to distinguish different objects, not to describe a specific order. Furthermore, the terms "comprising" and "having," and any variations thereof, are intended to cover non-exclusive inclusion. For example, a process, method, system, product, or apparatus that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or apparatuses.

[0035] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a separate or alternative embodiment mutually exclusive with other embodiments. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0036] The current practice of relying primarily on manual writing and experience-driven approaches to create product marketing copy results in low efficiency and high costs. Existing automated copy generation methods not only suffer from inconsistent copy quality, difficulty in maintaining brand consistency and compliance, but also struggle to effectively integrate product selling points with marketing copy, leading to fragmented marketing materials. Furthermore, with the rapid development of cross-border e-commerce, merchants need to efficiently produce high-quality marketing materials in multilingual, multi-national, and multicultural market environments. It is difficult to generate structured and controllable copy based on target country / site, style, and template requirements.

[0037] To address the aforementioned problems, this application provides a method and system for automatically generating product marketing copy. The embodiments of this application will be described in detail below with reference to the accompanying drawings.

[0038] Please see Figure 1 , Figure 1 This is a flowchart illustrating a method for automatically generating product marketing copy provided in this application. The method includes the following steps.

[0039] S101, obtain the product information of the target product input by the user at the selling point analysis node, as well as the competitor information of at least one set of competitors of the target product; and generate a structured selling point list at the selling point analysis node based on the product information and the competitor information.

[0040] The product information includes the target product title and description, as well as an image of the target product. Competitor information includes competitor titles and descriptions. The system can use web crawlers to search for competitor images and other related information on web pages based on this competitor information. In the selling point analysis module, core selling points can be extracted from the input product information and several competitor information entries, generating a structured selling point list. This structured selling point list includes fields such as the target product's core selling points, supplementary explanations of the core selling points, and image content plans for the core selling points. The image content field can be an array containing supported selling points and visual image content for each item. This image content field can be an array of length 3. For example, this structured selling point list may include more than 5 selling point entries and 3×N image plans. In practice, when generating marketing copy, product information and competitor information can be input by importing files. For example… Figure 2 As shown, in the "New Task" module, users create an automated marketing copy generation task by creating a task name, selecting a task workflow, and importing the corresponding file. When creating an automated generation task, as follows... Figure 3 As shown, users can choose the number of similar items to research, the number of virtuous items to research, the number of storylines, and the number of templates generated for storylines. Then, for example... Figure 4 As shown, based on the content selected by the user, the user can upload relevant file content by clicking the file selection control, thereby realizing the automated generation of tasks related to the target product.

[0041] Perform the following operations on any subsequent execution node:

[0042] S102, obtain the product supplementary information input by the user at the current execution node, and the node output results of at least one executed node, wherein the node output results of the at least one executed node include the structured selling point list generated by the selling point analysis node; generate the node output result of the current execution node based on the product supplementary information and the node output results of the at least one executed node; and determine the node output result of the current execution node as the reference output result if the user selects to jump to the next execution node, or if the current execution node is a copywriting generation node.

[0043] Each execution node indicates a workflow, which can be performed sequentially based on the order of the execution nodes to obtain reference output results. This workflow includes selling point analysis, image tag generation, storyline generation, style matching, template matching, and copywriting generation. Therefore, it corresponds to the following nodes in sequence: image tag generation node for outputting image tags, storyline generation node for generating the target product's storyline, style matching node for determining the style attributes of the target product, template matching node for determining the output template, and copywriting generation node for generating marketing copy. During actual execution, users can skip or rewind. That is, after entering each execution node, the user can choose whether to skip that node or return to a previously executed node for re-execution. Specifically, supplementary information from the current execution node can also serve as supplementary information for the next execution node. If the supplementary information for the next execution node has already been obtained during the execution of the previous execution node, then execution will proceed based on the information obtained from the previous execution node.

[0044] In the specific implementation, when each execution node is executed, if the supplementary information entered by the user is the same as the supplementary information entered in the past, and the output result of the executed node obtained by the execution node is the same as the output result of the past, then the current request is an idempotent request, and the output result of the current request and the corresponding current execution node in the past are cached.

[0045] In the specific implementation, at the image tag generation node, the supplementary information input by the user can include selling point attribute information, which is structured information used to supplement the details of core selling points. Then, based on the target product information, competitor information, selling point attribute information, and structured selling point list, the target product image and competitor images are obtained, and the target product image and competitor images are tagged. The tag content can include image type, implied selling points, image content summary, and quality score. For example, image types include white background images, size images, and scene images.

[0046] In practice, at the style matching node, supplementary information can include user-uploaded selling point attributes. Then, based on these attributes, a structured list of selling points, and the target product's storyline, the style attribute is determined from a pre-defined style library, matching the target product's core selling points and storyline. This style attribute includes style tags, tone parameters, and constraints. In other words, at the style matching node, the optimal style or style combination can be matched based on the style library (such as dynamic sports, modern minimalism, green living, etc.) and site / audience preferences.

[0047] In the specific implementation, at the template matching node, based on the selling point attribute information, selling point list, target product storyline and style matching results uploaded by the user, a corresponding number of templates can be selected from the merchant's custom template library. The templates contain placeholders and rendering rules, and the selected target's ID list and rendering metadata are output.

[0048] In one possible embodiment, before jumping to the next execution node according to the user's selection, the method further includes: determining whether the node output result of the current execution node has been obtained; if the node output result has not been obtained, determining whether the current execution node is a necessary execution node; if it is a necessary execution node, outputting a prompt message to indicate the necessity of executing the current execution node; after outputting the prompt message, if the user selects to jump to the next execution node again without obtaining the node output result, generating a default output result of the current execution node based on the node output results of at least one already executed node, and jumping to the next execution node.

[0049] Before navigating to the next execution node based on user selection, the method first uses the workflow engine's status check module to determine whether the node output of the current execution node has been obtained. If no output has been obtained, the node management unit is invoked to determine whether the node is a necessary execution node (i.e., a strongly dependent node). The determination of a necessary execution node is based on a predefined node dependency graph, which is stored using a Directed Acyclic Graph (DAG) structure. Each node contains a node type, a list of dependent edges, a list of child nodes, and a failure strategy configuration. For example, for a style matching node, if its output directly affects the core parameters of subsequent content generation nodes, it is marked as a necessary execution node.

[0050] When the current node is identified as a necessary execution node, the system outputs dynamic prompts through the user interface module. These prompts include potential business risks associated with skipping the node, such as decreased content quality, and reasons for recommending execution. The prompts can be generated by combining priority parameters from the node configuration with historical execution data, such as the error rate after skipping the node in the past, to enhance persuasiveness. If the user still chooses to skip after the prompts, the workflow engine triggers a default result generation process. First, it extracts the output of at least one executed node from the context cache; second, based on these results, it matches a predefined default template using the rules engine. For example, when skipping a style-matching node, the default style selection logic includes selecting "minimalist style" if the executed node contains "core selling point A". This matching process can be implemented using a weighted algorithm, with the weights determined by the type of the node's dependent edges. Figure 5As shown, the display interface includes storyline generation nodes, including a story-based insight module for analyzing product features and uncovering product storylines and selling points, and a storyline generation module for generating complete product storylines based on the insight results. There's also a template matching module for matching the most suitable marketing templates and copywriting frameworks, and a copywriting generation + template summary + compliance check and replacement module for generating marketing copy based on templates and styles, and providing summaries and optimization suggestions. Users can perform operations at corresponding nodes by entering different modules, and users can also skip nodes corresponding to a module. For example... Figure 5 In this mode, users can skip the template matching module and directly generate copy.

[0051] In the implementation, timeout and retry policies can be configured for each node to handle execution anomalies. The timeout threshold can be dynamically adjusted based on the node's historical execution duration. The retry policy uses an exponential backoff algorithm, with a default maximum of 3 retries. If a node ultimately fails to output a result due to timeout or retry failure, the system activates a rollback model. For data-driven nodes, the rollback model uses historical averages to fill the gaps; for logic processing nodes, the rollback model calls a simplified rule template to generate default output. After all default results are generated, an audit log is recorded, marked with "skipped generation" or "rollback generation," for subsequent quality traceability.

[0052] The above process utilizes the workflow engine's topology sorting algorithm to schedule nodes. When resolving node dependencies, the engine allows users to choose to skip non-strongly dependent nodes and enforce prompts and confirmations on necessary nodes, thereby balancing operational efficiency and system reliability.

[0053] In one possible embodiment, generating the default output result of the currently executing node based on the node output results of the at least one executed node includes: determining whether the node output result of the currently executing node depends on the supplementary information; if it depends on the supplementary information, generating default supplementary information based on the product information and the competitor information, and generating the default output result based on the node output results of the at least one executed node and the default supplementary information.

[0054] In the process of generating the default output of the currently executing node based on the output of at least one executed node, the system first uses a dependency analysis module to determine whether the output of the currently executing node depends on external supplementary information. If a dependency exists, the default supplementary information generation process is triggered. This process can be based on a combination of historical supplementary information input by the user in the executed nodes and network information obtained in real time through web crawling. When combining information, the purpose of supplementary information can be determined first, and weights can be assigned to historical supplementary information and network information respectively. Then, the historical supplementary information and network information are combined in a weighted manner based on the weights. The purpose of supplementary information can include two options: improving the accuracy of the output result and improving the coverage of the output result. For example, when the purpose of supplementary information is to improve the accuracy of the output result, the weight of historical supplementary information is higher than that of network information.

[0055] As can be seen, when the output depends on supplementary information, generating supplementary information based on product information and competitor information can automatically generate the node output when the user skips it, ensuring the smooth execution of candidate nodes.

[0056] S103, perform compliance verification on the reference output result according to the target market to which the target product belongs, and output the marketing copy corresponding to the reference output result after the verification is passed.

[0057] This process can incorporate self-assessment / peer assessment (LLM-based Criticism), regular expression / rule validation, terminology / sensitive word review, and compliance models for compliance review. When generating marketing copy, it can be based on selling point attributes, a list of selling points, style matching, template matching results, and target country / language. Specifically, this can include calling a Large Language Model (LLM) under FunctionSchema constraints to fill in placeholders, generating titles, selling points, image copy, and content suggestions. Then, based on the content template, the generated titles, selling points, image copy, and content suggestions are visually assembled into a Markdown preview, and a readable version for business review is output, ultimately resulting in Markdown format marketing copy. This marketing copy includes text and / or images.

[0058] As can be seen, in this embodiment, according to the product marketing copy automatic generation method and system provided in this application, a structured list of selling points is generated at the selling point analysis node, and then corresponding output results are generated based on user selections at different execution nodes such as image tag generation node, storyline generation node, style matching node, template matching node, and copy generation node. This achieves structured generation of product marketing copy for different target markets, improving the quality of copy generation. This ensures that the generated marketing copy not only effectively integrates with the core selling points of the product but also adapts to different target markets, improving the compliance of the copy.

[0059] In one possible embodiment, generating the target product storyline includes: obtaining supplementary product information input by the user at the storyline generation node, the supplementary product information including attribute information of the core selling points; generating a competitor storyline based on the attribute information and competitor image tags of at least one set of competitors output by the image tag generation node; and generating the target product storyline based on the attribute information, the structured selling point list, and the competitor storyline.

[0060] The specific process of generating a competitor storyline based on the attribute information and competitor image tags includes: First, determining a first target scenario based on the attribute information of the core selling points. This first target scenario clarifies the characteristics of the target audience, such as outdoor hiking enthusiasts aged 25-35, and the core usage scenario, such as short-to-medium distance hiking in the rain. Furthermore, the first target scenario matches the application scenario of the core selling points in the attribute information. For example, if the attribute information is "IPX7 waterproof rating, suitable for rainy outdoor scenarios," then the first target scenario is defined as "the scenario where outdoor hiking enthusiasts need waterproof equipment in rainy environments." Second, filtering image association information from the competitor image tags that matches the first target scenario. Finally, a cinematic approach can be used to weave the images corresponding to the filtered competitor image tags into a 5-act story structure to form a competitor storyline. Each act of the 5-act story structure is associated with the corresponding competitor image tag information.

[0061] As can be seen, this embodiment can improve the contextual relevance, text-image synergy, and narrative logic of competitor storylines, providing a precise benchmark for the differentiated construction of subsequent target product storylines, and effectively solving the problems of ambiguous context, fragmented text-image synergy, or chaotic narrative in existing technologies.

[0062] In one possible embodiment, generating a competitor storyline from multiple competitor images and competitor image tags for each competitor image, generated based on the attribute information and the image tags, includes: determining a first target scenario based on the attribute information; generating a multi-act story structure under the target scenario based on the multiple competitor images and the competitor image tags, wherein the multi-act story structure constitutes the competitor storyline, each act corresponding to a core selling point, the arguments and justifications for the core selling point, and the narrative techniques, image style, and emotional triggers of the core selling point.

[0063] Each scene's story structure corresponds to a core selling point, its supporting arguments and justifications, and includes the narrative techniques, style, and emotional triggers of that sequence of shots. In practice, the primary target scenario can be multiple; that is, multiple core selling points can each correspond to a target scenario, or multiple core selling points can correspond to the same target scenario. For example... Figure 6 As shown, the competitor is an air purifier, and the corresponding scenario is: newly renovated families seeking healthy breathing, concerned about formaldehyde and bacteria, and with high demands for their baby's safety, requiring a powerful air purifier to solve the environmental safety problems of moving into a new home. The first act's story structure corresponds to the argument, or core selling point: powerful formaldehyde removal, ensuring peace of mind when moving into a new home; the evidence is: 99% formaldehyde removal rate, quickly purifying a 120m² space; the proof is: ensuring safe air in the new home, giving the family greater peace of mind; the narrative technique is: combining data and promises to establish a reassuring atmosphere; the style is: Nordic minimalism, warm home; the emotional trigger is: health protection, anticipation of moving in. The second act's story structure corresponds to the argument: purification not only removes formaldehyde but also kills bacteria; the evidence is: 99% sterilization rate, practical for everyday use; the proof is: continuous purification, ensuring a healthy environment for the whole family; the narrative technique is: progressive function, comparative explanation; the style is: warm home; the emotional trigger is: health protection, parent-child resonance.

[0064] As can be seen, this embodiment can improve the ability of competitors' storylines to accurately convey core selling points, enhance narrative persuasiveness and user emotional resonance, while also increasing adaptability to multi-scenario marketing needs.

[0065] In one possible embodiment, generating the target product storyline based on the attribute information, the structured selling point list, and the competitor storyline includes: determining the target product's advantages against the competitor for each core selling point based on the attribute information and the structured selling point list; and generating the target product storyline based on the advantages and the competitor storyline.

[0066] In generating the target product storyline, multiple story structures can be generated separately. Each story structure still includes arguments and evidence for the core selling points, as well as the narrative techniques, visual style, and emotional triggers for those core selling points. It should be noted that each story structure of the target product can be based on the same story structure of a competitor's product with the same core selling points, with the advantages of this solution emphasized. Simultaneously, the narrative fluency and rationality of competitor storylines can be evaluated, and optimized content can be generated based on the evaluation results. Then, the target product storyline is generated by combining the optimized content, the advantages, and the competitor's storyline.

[0067] As can be seen, this embodiment, by using the same story structure based on the core selling points of competing products, emphasizes the advantages of the target product in each scene's narrative. This ensures the target product's storyline is aligned with the competitor's in terms of narrative framework, facilitating a direct comparison of differences between the two for users. It also precisely highlights the target product's advantages in its core selling points, avoiding ambiguity in the communication of these advantages. By evaluating the narrative fluency and rationality of the competitor's storyline and generating optimized content, potential logical gaps and abrupt transitions in the competitor's narrative can be addressed, making the target product's storyline more coherent and logically rigorous, thus enhancing user comprehension and trust. Simultaneously, each scene retains the arguments, evidence, appropriate narrative techniques, visual style, and emotional triggers for the core selling points, ensuring the target product's advantages are solidly supported and resonate with users emotionally. This further enhances the storyline's marketing appeal, laying the foundation for generating high-quality marketing copy based on this storyline, ultimately contributing to increased user recognition and conversion potential for the target product. This embodiment can improve the differentiated competitiveness, narrative persuasiveness, and overall quality of the target product's storyline, while ensuring benchmarking against competing product storylines and narrative logical coherence.

[0068] In one possible embodiment, the copy generation node performs the following operations: generates text information corresponding to each story structure based on the target product storyline; determines at least one target image corresponding to each story structure based on the text information and the image tags; and generates marketing copy information corresponding to each story structure based on the text information and the at least one target image.

[0069] The generated marketing copy and target image share the same story structure, thus aligning on the core selling points. Specifically, the marketing copy is also linked to the preferences and cognitive biases of the target market. For example, when the core selling point of the target product is "natural raw stone," the output text includes the corresponding language and preferences from that country, given that the target market is China. When the core selling point is "solid wood base," the corresponding image is an image highlighting the "solid wood" feature of the base. In other words, the images and text in this application share the same story structure, the selling points displayed in the images are consistent with the copy, and the image style is highly unified with the copy.

[0070] As can be seen, this embodiment ensures that the core selling points of the copy and the target image are completely consistent and the style is highly unified by placing them within the same story structure. This effectively solves the problems of fragmented text and images and misaligned selling point delivery in existing technologies, allowing users to quickly grasp the core value through the synergistic perception of textual descriptions and visual presentations, thus improving information reception efficiency. Simultaneously, the marketing copy is linked to the target market, adapting to the language habits and cognitive preferences of the target market, reducing localization communication costs, and enhancing the emotional identification and acceptance of local users. This design, which combines text and images with the same selling points and localization, also allows the marketing materials to form a unified narrative context, avoiding ambiguity in the delivery of core selling points or style confusion, further strengthening brand awareness consistency, and providing high-quality, highly adaptable material support for subsequent deployment in the target market, helping to improve the exposure and conversion rate of the product.

[0071] In one possible embodiment, determining at least one target image corresponding to each scene story structure based on the text information and the image tags includes: performing the following operations for each scene story structure: selecting a first group of images from reference images whose image tags include scene image types; selecting a second group of images from the first group of images whose product selling points, as included in the image tags, correspond to the target context; selecting a third group of images from the second group of images whose image content summary, as included in the image tags, includes key image elements, the key image elements including image elements associated with the core selling points; performing a quality score on the third group of images, and determining the images whose quality score is higher than a preset value as target images.

[0072] When filtering competitor images, priority is given to images tagged as scene images, excluding white-background images that only show the product's appearance. Then, the product selling point fields in the tags are confirmed to be highly relevant to core selling point attributes, such as waterproofing and lightweight. Simultaneously, image content summaries are extracted from the filtered competitor image tags to identify the actual competitor images. For example, "Competitor's rain jacket used in the rain; rainwater rolls off the fabric surface without penetration." Furthermore, images with quality scores meeting a preset acceptable threshold are selected to ensure the image information is clear and usable for storytelling.

[0073] As can be seen, this embodiment can improve the adaptability of the target image to the story structure of each scene, the relevance of the core selling points, and the reliability of image quality.

[0074] In one possible embodiment, the quality scoring of the third set of images includes: performing the following operations for each image in the third set of images: determining the image sharpness of the image and the relevance of the image to the core selling point; determining an image quality score based on the image sharpness and the relevance to the selling point; determining image risk based on the target market and the marketing copy placement platform; performing a safety quality score on the third set of images based on the image risk; and determining a final quality score for the image based on the image quality score and the safety quality score.

[0075] Regarding image sharpness determination, the system extracts key sharpness indicators from competitor images using image analysis algorithms, including but not limited to image resolution, noise rate, and detail retention. This detail retention can be correlated with visual details related to the core selling point. For example, if the core selling point is "waterproof," the image must clearly show the trajectory of water droplets rolling off the fabric surface without blurring or blurring. If the competitor image is a "close-up of the waterproof fabric" and includes the visual element of "water droplets rolling off the fabric without penetration," then the image quality is high.

[0076] The process of determining image risks based on the target market and the marketing copy placement platform can be achieved by constructing a risk identification matrix by combining the compliance rules of both. First, for the target market, a pre-defined compliance rule library for a specific country or region can be accessed. For example, if the target market is country A, the image needs to be checked for risks such as "missing CE certification mark" or "incorrect environmental labeling." Second, for the marketing copy placement platform, the platform's image rule library can be accessed. For instance, platform B prohibits images containing "absolute terms (such as 'best waterproof gear')," "infringing brand logos (such as unauthorized outdoor brand logos)," and "blurred watermarks or materials of unknown origin." Optical Character Recognition (OCR), logo comparison, and compliance terminology matching technologies can be used to scan competitor images for risks and obtain a safety and quality score. Finally, based on the principle of compliance priority, a higher weight can be assigned to the safety and quality score, and the final quality score can be obtained through a weighted summation method.

[0077] As can be seen, this embodiment can improve the accuracy and compliance of the third group of competitor image selection, and provide high-quality, low-risk image material support for subsequent shot generation and competitor storyline construction.

[0078] The following describes an automatic product marketing copy generation system provided in this application. The automatic product marketing copy generation system described below corresponds to the automatic product marketing copy generation method described above.

[0079] Please see Figure 7 The product marketing copy automatic generation system 800 includes a selling point analysis module 801, which is used to obtain product information of the target product input by the user at the selling point analysis node, as well as the competitor information of at least one set of competitors of the target product; and generate a structured selling point list based on the product information and the competitor information. The structured selling point list includes the core selling points of the target product, supplementary explanations of the core selling points, and image content plans for the core selling points.

[0080] Each subsequent execution module is used to perform the following operations: obtain supplementary product information input by the user at the current execution node, and the node output results of at least one executed module, wherein the node output results of the at least one executed module include a structured list of selling points generated by the selling point analysis module; generate the node output result of the current execution module based on the supplementary product information and the node output results of the at least one executed module; and determine the node output result of the current execution module as a reference output result based on the user's selection to jump to the next execution module, or when the current execution module is a copywriting generation module; each subsequent execution module sequentially includes an image tag generation module 802 for outputting image tags, a storyline generation module 803 for generating the target product storyline, a style matching module 804 for generating and determining the style attributes of the target product, a template matching module 805 for determining the output template, and the copywriting generation module 806 for generating marketing copy; a compliance verification module 807 is used to perform compliance verification on the reference output result according to the target market to which the target product belongs, and output the marketing copy corresponding to the reference output result after the verification is passed.

[0081] In one possible embodiment, the product marketing copy automatic generation system 800 further includes a node jump module. Before jumping to the next execution node according to the user's selection, the node jump module is used to: determine whether the node output result of the current execution node has been obtained; if the node output result has not been obtained, determine whether the current execution node is a necessary execution node; if it is a necessary execution node, output a prompt message, the prompt message being used to indicate the necessity of executing the current execution node; after outputting the prompt message, if the node output result has not been obtained again, and the user's instruction to jump to the next execution node is received, generate a default output result of the current execution node based on the node output results of at least one already executed node, and jump to the next execution node.

[0082] In one possible embodiment, in generating the default output result of the currently executing node based on the node output results of the at least one executed node, the node jump module is configured to: determine whether the node output result of the currently executing node depends on the supplementary information; if it depends on the supplementary information, generate default supplementary information based on the product information and the competitor information, and generate the default output result based on the node output results of the at least one executed node and the default supplementary information.

[0083] In one possible embodiment, in generating the target product storyline, the storyline generation module 803 is specifically configured to: obtain supplementary product information input by the user at the storyline generation node, the supplementary product information including attribute information of the core selling points; generate a competitor storyline based on the attribute information and competitor image tags of at least one set of competitors output by the image tag generation node; and generate the target product storyline based on the attribute information, the structured selling point list, and the competitor storyline.

[0084] In one possible embodiment, in generating a competitor storyline from multiple competitor images and competitor image tags for each competitor image output by the generation node based on the attribute information and the image tag, the storyline generation module 803 is specifically configured to: determine a first target scenario based on the attribute information; generate a multi-act story structure under the target scenario based on the multiple competitor images and the competitor image tags, wherein the multi-act story structure constitutes the competitor storyline, and each act of the story structure corresponds to a core selling point, the arguments and justifications of the core selling point, and the narrative techniques, image style, and emotional triggers of the core selling point.

[0085] In one possible embodiment, in generating the target product storyline based on the attribute information, the structured selling point list, and the competitor storyline, the storyline generation module 803 is specifically configured to: determine the target product's advantages against the competitor at each core selling point based on the attribute information and the structured selling point list; and generate the target product storyline based on the advantages and the competitor storyline.

[0086] In one possible embodiment, the copywriting generation module 806 is specifically used to: generate text information corresponding to each scene story structure based on the target product storyline; determine at least one target image corresponding to each scene story structure based on the text information and the image tags; and generate marketing copy information corresponding to each scene story structure based on the text information and the at least one target image.

[0087] In one possible embodiment, regarding the determination of at least one target image corresponding to each scene story structure based on the text information and the image tags, the copywriting generation module 806 is specifically configured to: perform the following operations for each scene story structure: select a first group of images from reference images whose image tags include scene image types; select a second group of images from the first group of images whose product selling points, as included in the image tags, correspond to the target context based on the target context; select a third group of images from the second group of images whose image content summary, as included in the image tags, includes key image elements, the key image elements including image elements associated with the core selling points; perform a quality score on the third group of images, and determine the images whose quality score is higher than a preset value as target images.

[0088] In one possible embodiment, regarding the quality scoring of the third set of images, the copywriting generation module 806 is specifically configured to: perform the following operations for each image in the third set of images: determine the image clarity and the relevance of the image to the core selling point; determine an image quality score based on the image clarity and the relevance of the selling point; determine the image risk based on the target market and the marketing copywriting platform; perform a safety quality score on the third set of images based on the image risk; and determine the final quality score of the image based on the image quality score and the safety quality score.

[0089] Please see Figure 8 , Figure 8 This is a schematic diagram of the structure of an electronic device provided in this application. For example... Figure 8 As shown, the electronic device may include a processor 910, a communications interface 920, a memory 930, and a communication bus 940. The processor 910, communications interface 920, and memory 930 communicate with each other via the communication bus 940. The processor 910 can call logical instructions stored in the memory 930 to execute the aforementioned method for automatically generating product marketing copy.

[0090] Furthermore, the logical instructions in the aforementioned memory 930 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0091] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to perform the product marketing copy automatic generation method provided in the above embodiments.

[0092] This application also provides a computer program product, including a computer program that, when executed by a processor, implements any of the above-described methods for automatically generating product marketing copy.

[0093] The above embodiments can be implemented, in whole or in part, by software, hardware, firmware, or any other combination thereof. When implemented using software, the above embodiments can be implemented, in whole or in part, as a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, all or part of the processes or functions described in the embodiments of this application are generated. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions can be stored in a computer-readable storage medium or transmitted from one computer-readable storage medium to another. For example, the computer instructions can be transmitted from one website, computer, server, or data center to another website, computer, server, or data center via wired or wireless means. The computer-readable storage medium can be any available medium that a computer can access or a data storage device such as a server or data center that includes one or more sets of available media. The available medium can be a magnetic medium (e.g., floppy disk, hard disk, magnetic tape), an optical medium (e.g., DVD), or a semiconductor medium. A semiconductor medium can be a solid-state drive.

[0094] This application also provides a computer storage medium storing a computer program for electronic data interchange, which causes a computer to perform some or all of the steps of any of the methods described in the above method embodiments, wherein the computer includes an electronic device.

[0095] It should be noted that, for the sake of simplicity, the foregoing method embodiments are all described as a series of actions. However, those skilled in the art should understand that this application is not limited to the described order of actions, as some steps may be performed in other orders or simultaneously according to this application. Furthermore, those skilled in the art should also understand that the embodiments described in the specification are preferred embodiments, and the actions and modules involved are not necessarily essential to this application.

[0096] In the above embodiments, the descriptions of each embodiment have different focuses. For parts not described in detail in a certain embodiment, please refer to the relevant descriptions in other embodiments.

[0097] In the several embodiments provided in this application, it should be understood that the disclosed apparatus can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical or other forms.

[0098] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0099] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0100] If the aforementioned integrated units are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage device (CMD). Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a memory and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned memory includes various media capable of storing program code, such as USB flash drives, read-only memory (ROM), random access memory (RAM), portable hard drives, magnetic disks, or optical disks.

[0101] Those skilled in the art will understand that all or part of the steps in the various methods of the above embodiments can be implemented by a program instructing related hardware. The program can be stored in a computer-readable storage device, which may include: a flash drive, a read-only memory (ROM), a random access memory (RAM), a magnetic disk, or an optical disk, etc.

[0102] The embodiments of this application have been described in detail above. Specific examples have been used to illustrate the principles and implementation methods of this application. The description of the above embodiments is only for the purpose of helping to understand the method and core ideas of this application. At the same time, for those skilled in the art, there will be changes in the specific implementation methods and application scope based on the ideas of this application. Therefore, the content of this specification should not be construed as a limitation of this application.

Claims

1. A method for automatically generating product marketing copy, characterized in that, include: Obtain product information of the target product input by the user at the selling point analysis node, as well as competitor information of at least one set of competitors of the target product; And at the selling point analysis node, a structured selling point list is generated based on the product information and the competitor information. The structured selling point list includes the core selling points of the target product, supplementary explanations of the core selling points, and image content plans for the core selling points. For any subsequent execution node, perform the following operations: obtain the product supplementary information input by the user at the current execution node, and the node output results of at least one executed node, wherein the node output results of at least one executed node include the structured selling point list generated by the selling point analysis node; The node output result of the current execution node is generated based on the product supplementary information and the node output results of at least one executed node; The system will jump to the next execution node based on the user's selection, or, if the current execution node is a copywriting generation node, determine the node output result of the current execution node as the reference output result; the subsequent execution nodes are, in sequence, an image tag generation node for outputting image tags, a storyline generation node for generating the target product storyline, a style matching node for generating and determining the style attributes of the target product, a template matching node for determining the output template, and the copywriting generation node for generating marketing copy; The reference output results are validated for compliance based on the target market to which the target product belongs, and the marketing copy corresponding to the reference output results is output after the validation is passed.

2. The method according to claim 1, characterized in that, Before jumping to the next execution node based on the user's selection, the method further includes: Determine whether the node output result of the currently executing node has been obtained; If the output result of the node is not obtained, determine whether the currently executing node is a necessary execution node; If it is a necessary execution node, a prompt message is output, which is used to indicate the necessity of the execution of the current execution node; After outputting the prompt information, if the user selects to jump to the next execution node without obtaining the node output result again, the default output result of the current execution node is generated based on the node output result of the at least one already executed node, and the user jumps to the next execution node.

3. The method according to claim 2, characterized in that, The step of generating the default output result of the currently executing node based on the node output results of the at least one executed node includes: Determine whether the node output of the currently executing node depends on the supplementary information; If the supplementary information is relied upon, then default supplementary information is generated based on the product information and the competitor information, and the default output result is generated based on the node output result of the at least one executed node and the default supplementary information.

4. The method according to any one of claims 1-3, characterized in that, The generation of the target product storyline includes: Obtain the product supplementary information input by the user at the storyline generation node, the product supplementary information including the attribute information of the core selling point; A competitor storyline is generated based on the attribute information and the competitor image tags of at least one set of competitors output by the image tag generation node. The target product storyline is generated based on the attribute information, the structured selling point list, and the competitor storylines.

5. The method according to claim 4, characterized in that, The step of generating a competitor storyline based on the attribute information and the image tag generation node outputs multiple competitor images and competitor image tags for each competitor image includes: The first target scenario is determined based on the attribute information; Based on the multiple competitor images and their tags, a multi-scene story structure is generated in the target context. The multi-scene story structure constitutes the competitor storyline. Each scene story structure corresponds to a core selling point, the arguments and justifications for the core selling point, and the narrative techniques, image style, and emotional triggers of the core selling point.

6. The method according to claim 5, characterized in that, The step of generating the target product storyline based on the attribute information, the structured selling point list, and the competitor storylines includes: Based on the attribute information and the structured list of selling points, determine the advantages of the target product against the competitors in each core selling point; The target product storyline is generated based on the advantages and the competitor storylines.

7. The method according to claim 5, characterized in that, Perform the following operations on the text generation node: Generate text information corresponding to each scene's story structure based on the target product's storyline; Based on the text information and the image tags, at least one target image corresponding to each scene's story structure is determined; Marketing copy information corresponding to each scene's story structure is generated based on the text information and the at least one target image.

8. The method according to claim 7, characterized in that, Determining at least one target image corresponding to each scene's story structure based on the text information and the image tags includes: Perform the following operations for each scene's story structure: Select the first group of images whose image type is scene graph from the reference images; Based on the target context, select a second set of images from the first set of images whose product selling points are included in the image tags and correspond to the target context; From the second set of images, select the image content summary of the image tag as a third set of images that includes key image elements, the key image elements including image elements associated with the core selling point; The third group of images is scored for quality, and the images with quality scores higher than a preset value are identified as target images.

9. The method according to claim 8, characterized in that, The quality scoring of the third group of images includes: Perform the following operations on each image in the third group of images: Determine the image clarity and the relevance of the image to the core selling point; An image quality score is determined based on the image sharpness and the relevance to the selling points; Image risks are determined based on the target market and marketing copy placement platforms. The third group of images is scored for safety quality based on the image risk. The final quality score of the image is determined based on the image quality score and the security quality score.

10. A system for automatically generating product marketing copy, characterized in that, include: The selling point analysis module is used to obtain product information of the target product input by the user in the selling point analysis node, as well as the competitor information of at least one set of competitors of the target product; A structured list of selling points is generated based on the product information and the competitor information. The structured list of selling points includes the core selling points of the target product, supplementary explanations of the core selling points, and a plan for image content related to the core selling points. Each subsequent execution module is used to perform the following operations: obtain supplementary product information input by the user at the current execution node, and the node output results of at least one executed module, wherein the node output results of the at least one executed module include a structured list of selling points generated by the selling point analysis module; generate the node output result of the current execution module based on the supplementary product information and the node output results of the at least one executed module; The system will then jump to the next execution module based on the user's selection, or, if the current execution module is the copywriting generation module, determine the node output result of the current execution module as the reference output result; the subsequent execution modules are, in sequence, an image tag generation module for outputting image tags, a storyline generation module for generating the target product storyline, a style matching module for generating and determining the style attributes of the target product, a template matching module for determining the output template, and the copywriting generation module for generating marketing copy; The compliance verification module is used to perform compliance verification on the reference output result according to the target market to which the target product belongs, and output the marketing copy corresponding to the reference output result after the verification is passed.

Images