AI instruction structured design method and system based on 5W3H framework and storage medium

Abstract

The invention discloses an AI instruction structured design method and system based on a 5W3H framework and a storage medium, belongs to the technical field of artificial intelligence, and particularly relates to an AI instruction generation technology. The method comprises the following steps: receiving preliminary information input by a user based on a 5W3H framework through a user interface; calling at least one preset AI model, enhancing or expanding the initial information input by the user, and generating more accurate element content; combining all dimension information directly input by a user or subjected to AI enhancement into a structured AI instruction according to a preset rule; and outputting the structured AI instruction for a user to use. By introducing the structured 5W3H framework and an optional AI auxiliary enhancement mechanism, the threshold of designing a high-quality AI instruction by a user can be remarkably reduced, the accuracy and efficiency of AI generation content are improved, and the interactive experience of the user and the AI is improved.

Description

Technical Field

[0001] This application belongs to the field of artificial intelligence technology, and specifically relates to an AI instruction structured design method, system, and storage medium based on the 5W3H framework. Background Art

[0002] In recent years, with the rapid development of artificial intelligence technologies such as deep learning, the capabilities of AI content generation models, such as large language models (LLMs) and image generation models, have been significantly enhanced. They have shown great application potential in a variety of fields, including text creation, code generation, image design, and intelligent question-answering. Users guide these AI models to generate the desired content by providing instructions (i.e., prompts).

[0003] However, in the existing technology, users often face many challenges when designing AI instructions. First, the design of high-quality AI instructions usually relies on the user's experience and intuition, lacking systematic methodological guidance, making it difficult for ordinary users to quickly master them. Second, the instructions provided by users often have problems such as arbitrariness, ambiguity, incomplete information, or missing key elements. For example, the specific requirements of the task, the target audience, the expected style, or the output format may not be clear. These low-quality instructions directly affect the AI ​​model's understanding of the user's intent, which in turn causes the generated content to deviate from expectations and be of low quality. It may even require users to make multiple attempts and modifications, reducing the efficiency of AI applications and user experience.

[0004] While some research and practice have attempted to help users by providing instruction templates or examples, these approaches are often targeted at specific scenarios or models, lack universal applicability, and fail to fundamentally address the problem of systematically constructing comprehensive, clear, and unambiguous AI instructions. Users still need to spend a significant amount of time thinking through and organizing each aspect of the instruction.

[0005] Therefore, there is an urgent need for a method and system that can systematically guide users to design high-quality AI instructions, so as to lower the user's usage threshold, improve the quality and efficiency of AI-generated content, and enhance the controllability and predictability of AI interactions. Summary of the Invention

[0006] In response to the above-mentioned defects of the prior art, the present application provides a structured design method, system and storage medium for AI instructions based on the 5W3H framework. Through the structured 5W3H framework, it provides users with a clear and systematic AI instruction design process. Even if users lack professional experience, they can think comprehensively and construct high-quality instructions according to the guidance, significantly lowering the threshold for using AI tools.

[0007] In order to achieve the above object, the present invention adopts the following technical solutions: In a first aspect, the present invention proposes a structured design method for AI instructions based on the 5W3H framework, comprising the following steps: S1: Obtain user demand information; the user demand information is preliminary information input based on the 5W3H framework; S2: Generate corresponding AI instructions based on the user demand information; The AI ​​instruction also includes an emotional tone adapted to a specific application scenario. The instruction (Prompt) refers to a prompt word.

[0008] Preferably, before S2, the method further includes: calling at least one preset artificial intelligence model to perform multi-dimensional processing on the preliminary information to generate element content; The user demand information is optimized according to the element content.

[0009] The multi-dimensional processing of the preliminary information specifically includes: The preliminary information is processed in terms of semantic depth mining, logical relationship sorting, knowledge fusion and expansion, emotion and style, and uncertainty processing.

[0010] The preset artificial intelligence model includes a third-party large model or a locally deployed dedicated model.

[0011] The preliminary information input based on the 5W3H framework specifically includes: Mission: The specific task, topic, or problem that the AI ​​should address. Purpose: Explain the purpose, intent, or motivation of the content; Audience: Specify the relevant audience, personas, or stakeholders for the content, and the perspective the AI ​​should adopt; Time: establishing a time frame, historical context, or time-sensitive factors; Location: defines the geographical, cultural, or environmental context; Quantization parameters: set range, length, quantity or other quantization parameters; Implementation: Outline the desired approach, format, or structure; and Emotional Tone: Specify the emotional tone, style, or mood.

[0012] The step S2 generates corresponding AI instructions based on the user demand information, and further includes: Sorting the optimized user demand information according to a preset rule, and selectively inserting preset conjunctions or separators between each dimension information; The preset rules include at least the importance of information, keyword guidance, natural language connection and structured tags.

[0013] In a second aspect, the present invention further provides a system for a structured design method of AI instructions based on a 5W3H framework, the system comprising: An information receiving module is used to obtain user demand information; the user demand information is preliminary information input based on the 5W3H framework; An instruction combination module, configured to combine the preliminary information of each dimension input by the user into a structured AI instruction according to predetermined rules; wherein the AI ​​instruction also includes an emotional tone adapted to a specific application scenario; The instruction output module is used to output the structured AI instructions.

[0014] The system further comprises: The element enhancement module is arranged between the information receiving module and the instruction combination module, and is used to call at least one preset artificial intelligence model, perform multi-dimensional processing on the preliminary information, generate element content, and pass the element content to the instruction combination module to optimize the user demand information.

[0015] In a third aspect, a computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements a structured design method for AI instructions based on the 5W3H framework as described in the first aspect.

[0016] In a fourth aspect, a computer device comprises a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, the method for structured design of AI instructions based on the 5W3H framework as described in the first aspect is implemented.

[0017] This invention adopts the 5W3H framework to collect preliminary information input by users, covering dimensions such as execution tasks, purpose, audience, time, location, quantitative parameters, execution methods and emotional tone, making user demand information more comprehensive and structured, laying the foundation for generating accurate AI instructions.

[0018] Compared with the prior art, the present invention has the following advantages: Through the structured 5W3H framework, users are provided with a clear and systematic AI instruction design process. Even if users lack professional experience, they can think comprehensively and build high-quality instructions according to the guidance, significantly lowering the threshold for using AI tools.

[0019] The 5W3H framework described in the present invention can effectively guide users to provide detailed information from multiple dimensions, ensuring that the generated instruction content is comprehensive, specific, and context-rich, reducing ambiguity and information omissions, thereby improving the accuracy of the AI ​​model's understanding of user intentions. High-quality instructions can guide the AI ​​model to generate content that is more accurate, more relevant, and more in line with the user's multi-dimensional needs, significantly improving the quality and satisfaction of the final output. By providing more complete instructions at one time, the number of times users need to repeatedly modify and regenerate instructions due to poor instructions is reduced, the efficiency of human-computer interaction is improved, and time and computing resources are saved. The 5W3H framework has good versatility and can be applied to various types of AI content generation tasks (such as text, images, code, etc.) and different AI models. BRIEF DESCRIPTION OF THE DRAWINGS

[0020] Figure 1 This is a flow chart of an AI instruction structured design method based on the 5W3H framework in one embodiment.

[0021] Figure 2 This is a system diagram of an AI instruction structured design method based on the 5W3H framework in one embodiment.

[0022] Figure 3 This is a 5W3H instruction designer interface in one embodiment. DETAILED DESCRIPTION

[0023] To make the purpose, technical solutions, and advantages of the embodiments of this application more clear, the technical solutions will be clearly and completely described below in conjunction with the embodiments of this application. Obviously, the described embodiments are part of the embodiments of this application, not all of them. Based on the embodiments in this application, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of this application.

[0024] Example 1, as Figure 1 As shown, the present invention proposes a structured design method for AI instructions based on the 5W3H framework, comprising the following steps: S1: Obtain user demand information; the user demand information is preliminary information input based on the 5W3H framework; The user accesses a specially designed user interface (e.g. Figure 3 The interface clearly lists the eight dimensions of the 5W3H framework and provides input areas for each dimension. The eight dimensions (5W3H) and their guiding descriptions are as follows: Figure 3 As shown: The first dimension: What (required): Here, the user clearly defines the specific task the AI ​​needs to perform, the topic to explore, or the core problem to be solved. For example, "Write an analytical report on the future development trends of renewable energy."

[0025] The second dimension is “Why”: The user explains the purpose, intention, or desired outcome of generating the content. For example, “for internal company strategy discussions, helping decision-makers understand industry trends.”

[0026] The third dimension, “Who,” specifies the target audience, relevant roles, or stakeholders of the content, as well as the perspective or persona the AI ​​should adopt when generating content. For example, “The target audience is senior company executives who have a business background but may lack deep technical knowledge. The AI ​​should write from the professional perspective of an industry analyst.”

[0027] The fourth dimension: When: Users define the timeframe, historical context, timeliness requirements, or deadlines for the content. For example, “Analyze development trends for the next 5-10 years, and the report must be completed by the end of this week.”

[0028] The fifth dimension, “Where,” defines the user's geographic location, cultural background, platform environment, or specific application scenario for the content. For example, “focus on the Chinese market, while also taking into account the situation of major global economies. The reporting style should be suitable for formal business presentations.”

[0029] The sixth dimension, How Much, specifies the scope, length, quantity, level of detail, or other quantitative parameters of the content. For example, "The report should be approximately 3,000 words long and include at least three key trends, each supported by two to three specific case studies."

[0030] The seventh dimension, How to Do, outlines the methods, steps, format, structure, or specific technical requirements that users expect to be followed in content production. For example, "The report structure should include an introduction, analysis of trends, risks and opportunities, and conclusions and recommendations. Data should be cited from the latest releases of authoritative organizations, and key data should be presented in charts and graphs."

[0031] The eighth dimension, How Feel, specifies the emotional tone, writing style, voice, or emotional overtones that the content should convey. For example, "The report should maintain an objective, rigorous, and professional style while expressing cautious optimism about future development trends."

[0032] Users fill in the corresponding input boxes with specific information for these dimensions based on their needs. The "What" dimension is usually required, while the other dimensions are optional based on actual needs. The more complete the information, the higher the quality of the generated instructions.

[0033] Preferably, before S2, the method further includes: calling at least one preset artificial intelligence model to perform multi-dimensional processing on the preliminary information to generate element content; The user demand information is optimized according to the element content.

[0034] Preferably, the multi-dimensional processing of the preliminary information specifically includes: The preliminary information is processed in terms of semantic depth mining, logical relationship sorting, knowledge fusion and expansion, emotion and style, and uncertainty processing.

[0035] The preset artificial intelligence model includes a third-party large model API call or a locally deployed dedicated model.

[0036] Among them, the optional third-party large model API calls include but are not limited to: GPT series: Suitable for applications such as assisting with writing articles, stories, and poems. For example, it can generate product descriptions and advertising copy for marketers. It can also provide users with information query, scheduling, and reminder services, such as being used as a voice assistant in smart speakers.

[0037] Claude series: Suitable for extracting summaries and analyzing content from long research reports and news articles. For example, in fields such as law and finance, it helps check whether documents comply with relevant regulations and standards.

[0038] Deepseek: Applied to e-commerce platforms, search engines, and data mining, it can enhance user experience and help companies better recommend content, deliver targeted advertisements, and conduct data analysis.

[0039] Alibaba Cloud's Qwen: An AI product based on large-scale pre-trained models, it focuses on natural language processing tasks. It can be applied to a variety of tasks, including automated writing, dialogue generation, text summarization, and sentiment analysis. It is widely used in scenarios such as intelligent customer service, intelligent assistants, and semantic search, helping enterprises improve operational efficiency and enhance user interaction experiences.

[0040] Locally deployed large models refer to open source generative AI large models that can be deployed on the user's local system, such as Deepseek's series of models 1.5b, 7b, 8b, 14b, 32b, 70b, and 671b, Alibaba Cloud's 0.6b, 1.7b, 4b, 8b, 14b, 30b, 32b, and 235b, as well as other open source, locally deployable large models such as gemma3.

[0041] Specifically, after the user submits preliminary information, the AI ​​decomposition of the 5W3H is triggered. The system can selectively activate one or more pre-set AI models (which can be third-party large-scale model APIs such as the GPT series and Claude series, or locally deployed dedicated models) to analyze, understand, refine, expand, or optimize some or all dimensions of the user input. This step aims to leverage AI capabilities to compensate for any unclear expressions, insufficient information, or incomplete thinking, thereby improving the quality of each instruction component.

[0042] For example, if the user only enters "student" in the "Who" dimension, the AI ​​enhancement module can analyze other dimension information (such as "what" is "explaining introductory concepts of quantum physics"), and then enrich "student" to "high school students or junior college undergraduates who are interested in physics but lack basic knowledge and need easy-to-understand language and vivid examples."

[0043] For example, if a user mentions "using a clear structure" in "How to do", the AI ​​enhancement module can specify it based on the "what" being "product manual" as "adopting a chapter structure of 'product overview - features - how to use - precautions - troubleshooting', using subheadings for each section and illustrations for key steps."

[0044] This step is a key innovation of the present invention, as it uses AI to improve the quality of instruction elements, thereby indirectly improving the overall effectiveness of the final instruction. The enhanced information will replace or supplement the user's original input for subsequent instruction combinations.

[0045] S2: Generate corresponding AI instructions based on the user demand information; Among them, the AI ​​instructions also include an emotional tone adapted to the specific application scenario.

[0046] Preferably, the step S2 generates corresponding AI instructions according to the user demand information, and further includes: Sorting the optimized user demand information according to a preset rule, and selectively inserting preset conjunctions or separators between each dimension information; Furthermore, the system collects the 5W3H elements directly entered by the user or after AI enhancement. Then, according to a set of preset logic and formatting rules, these elements are organically combined to form a complete, comprehensive, and structured AI instruction. These preset rules include at least the importance of information, keyword guidance, natural language connections, and structured markup.

[0047] Importance of information: For example, arrange in a fixed order such as What, Why, Who, When, Where, How much, How to do, How feel.

[0048] Keyword guidance: Add the keywords of each dimension before the information of each dimension, such as "Task: [What content]; Purpose: [Why content];..." Natural language connection: Use connecting words and phrases to string together information from different dimensions into one or more natural descriptive texts.

[0049] Structured markup: Organize information of each dimension into JSON, XML, or other custom structured data formats to facilitate AI model parsing.

[0050] For example, a simple combination method might be to concatenate the contents of all non-empty dimensions into a long string in the form of "dimension name: dimension content", separated by semicolons or newlines.

[0051] More advanced combination methods may dynamically adjust the combination logic according to the characteristics of different AI models to generate an instruction format optimized for a specific model.

[0052] The "Generate Instructions" button on the interface (such as Figure 3 ) usually triggers the execution of this step.

[0053] The resulting structured AI instructions are then displayed to users through a user interface. Users can easily view and copy the instructions and then paste them into the target AI content generation tool (such as a chatbot interface, an AI writing assistant, or an input box in an image generation platform).

[0054] Through the method of this embodiment, users can systematically and efficiently create high-quality AI instructions that are rich in information and have a clear structure, thereby significantly improving the quality and relevance of AI-generated content.

[0055] Example 2, as Figure 2 As shown, the present invention also provides a system for the structured design method of AI instructions based on the 5W3H framework. The system can be a web application deployed in the cloud or a locally installed desktop or mobile application. The system includes: The information receiving module is used to obtain user demand information; the user demand information is preliminary information input based on the 5W3H framework; Figure 3 The "5W3H Instruction Designer" interface shown in the figure below contains: text input boxes or selection controls corresponding to the eight dimensions of What, Why, Who, When, Where, How Much, How to Do, and How Feel, for users to enter preliminary information; a "Generate Instruction" button, which triggers subsequent processing; and an area for displaying the final structured AI instruction. After the user completes the information for each dimension on the interface and clicks the "Generate Instruction" button, this module is responsible for collecting all the preliminary information entered by the user and passing it to subsequent modules for processing.

[0056] The instruction assembly module receives preliminary information from the information reception module (if the element enhancement module is not enabled or has not yet processed it) or the enhanced element content from the element enhancement module. The instruction assembly module internally defines one or more sets of instruction assembly rules. Based on the selected rules, it logically arranges and formats the information from each dimension to generate the final structured AI instruction.

[0057] The instruction output module receives the structured AI instructions generated by the instruction combination module 104 and passes them to the user interface module 101 to display them to the user in a designated area.

[0058] The system also includes: an element enhancement module, which is arranged between the information receiving module and the instruction combination module, and internally integrates or calls one or more AI models through an API. It can be a general LLM or a model optimized for a specific field. The element enhancement module is used to call at least one preset artificial intelligence model to perform multi-dimensional processing on the preliminary information, and according to the preset logic, for example, enhance all dimensions, or only process the dimensions specified by the user or determined by the system to need to be enhanced, to generate more accurate, richer or more contextual element content. And pass the element content to the instruction combination module to optimize the user demand information. Preferably, if the system configuration does not enable this module, the preliminary information obtained by the information receiving module will be directly passed to the instruction combination module.

[0059] Preferably, the system also includes: a control module, which is responsible for coordinating the operation processes of various modules within the system, for example, receiving user operation instructions, calling information reception, element enhancement (if enabled), instruction combination, instruction output and other modules in sequence, and processing data transmission between modules.

[0060] Preferably, the system also includes: an AI model interface module; if the element enhancement module is enabled, the AI ​​model interface module is responsible for communicating with an external or internal AI model, sending preliminary information to be enhanced, and receiving the enhancement results returned by the AI ​​model.

[0061] Preferably, the system further includes: a configuration and rule library; storing system configuration information, such as whether to enable the element enhancement module, the default AI model selection, instruction combination rules, etc. Administrators or advanced users can manage these configurations.

[0062] For further explanation, the workflow of the system of the present invention is as follows: The user enters preliminary information of each dimension of 5W3H through the user interface module and clicks "Generate Instructions".

[0063] The information receiving module collects this preliminary information.

[0064] Optionally, the control module calls the element enhancement module, which interacts with the AI ​​model through the AI ​​model interface to enhance the preliminary information.

[0065] The instruction combination module obtains preliminary information or enhanced element content and generates structured AI instructions based on preset rules.

[0066] The instruction output module passes the generated instructions to the user interface module for display.

[0067] The user interface of this system (such as Figure 3 The "5W3H Instruction Designer" is intuitive and easy to use. Users simply follow the instructions to obtain high-quality AI instructions. For example, after completing all required information in the "5W3H Instruction Designer," clicking "AI Parse 5W3H" and "Generate Instruction" triggers the integration of various dimensions into a complete instruction. The system automatically completes the aforementioned process and displays the completed AI instruction at the bottom of the interface (or in a designated area). Specifically, the system integrates content from eight dimensions to form a complete AI instruction. Users can use this AI instruction to generate the desired results in other AI applications. This AI instruction can be modified and reused multiple times.

[0068] Among them, clicking the "AI Analysis 5W3H" button will trigger the call of the large model to generate or optimize content in each dimension.

[0069] In a third embodiment, a computer-readable storage medium stores a computer program. When executed by a processor, the computer program implements the AI ​​instruction structured design method based on the 5W3H framework as described in the first aspect. The computer-readable storage medium can be any physical device capable of storing program code, such as ROM, RAM, a hard drive, a solid-state drive, a USB flash drive, or an optical disk (CD, DVD, etc.).

[0070] A fourth embodiment is a computer device comprising a processor, a memory, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it can implement the AI ​​instruction structured design method based on the 5W3H framework as described in Example 1, or constitute the system as described in Example 2. The computer device can be a personal computer, a server, a mobile device (such as a smartphone or tablet), an embedded system, or the like.

[0071] In summary, this invention provides a systematic and efficient AI instruction design method and system by introducing an innovative 5W3H framework and an optional AI-assisted enhancement mechanism. This method not only significantly lowers the barrier for users to design high-quality AI instructions and improves the accuracy and efficiency of AI-generated content, but also enhances the controllability and predictability of user-AI interactions in a structured manner, thus possessing broad application prospects and significant practical value.

[0072] Although example embodiments have been described herein with reference to the accompanying drawings, it should be understood that the above example embodiments are merely illustrative and are not intended to limit the scope of the present application. Various changes and modifications may be made therein by those skilled in the art without departing from the scope and spirit of the present application. All such changes and modifications are intended to be included within the scope of the present application as required by the appended claims.

[0073] It should be noted that, in this document, relational terms such as first and second, etc., are used only to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply the existence of any such actual relationship or order between these entities or operations. Moreover, the terms "comprises," "comprising," or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article, or device comprising a series of elements includes not only those elements, but also other elements not explicitly listed, or elements inherent to such process, method, article, or device. In the absence of further limitations, an element defined by the phrase "comprising a ..." does not exclude the presence of other identical elements in the process, method, article, or device comprising the element.

[0074] Although the present application is described in conjunction with the above specific embodiments, it is obvious that those skilled in the art can make many substitutions, modifications and variations based on the above content. Therefore, all such substitutions, modifications and variations are included within the spirit and scope of the appended claims.

Claims

1. A structured design method for AI instructions based on the 5W3H framework, characterized in that: Including steps: S1: Obtain user demand information; the user demand information is preliminary information input based on the 5W3H framework; S2: Generate corresponding AI instructions based on the user demand information; The AI ​​instructions include an emotional tone adapted to a specific application scenario.

2. The AI ​​instruction structured design method based on the 5W3H framework according to claim 1 is characterized in that: Before S2, it also included: Calling at least one preset artificial intelligence model to perform multi-dimensional processing on the preliminary information to generate element content; The user demand information is optimized according to the element content.

3. The AI ​​instruction structured design method based on the 5W3H framework according to claim 2 is characterized in that: The multi-dimensional processing of the preliminary information specifically includes: The preliminary information is processed in terms of semantic depth mining, logical relationship sorting, knowledge fusion and expansion, emotion and style, and uncertainty processing.

4. The AI ​​instruction structured design method based on the 5W3H framework according to claim 3 is characterized in that: The preset artificial intelligence model includes a third-party large model or a locally deployed dedicated model.

5. The AI ​​instruction structured design method based on the 5W3H framework according to claim 4 is characterized in that: The preliminary information input based on the 5W3H framework specifically includes: Mission: The specific task, topic, or problem that the AI ​​should address. Purpose: Explain the purpose, intent, or motivation of the content; Audience: Specify the relevant audience, personas, or stakeholders for the content, and the perspective the AI ​​should adopt; Time: establishing a time frame, historical context, or time-sensitive factors; Location: defines the geographical, cultural, or environmental context; Quantization parameters: set range, length, quantity or other quantization parameters; Implementation: Outline the desired approach, format, or structure; and Emotional Tone: Specify the emotional tone, style, or mood.

6. The AI ​​instruction structured design method based on the 5W3H framework according to claim 5 is characterized in that: The step S2 generates corresponding AI instructions based on the user demand information, and further includes: Sorting the optimized user demand information according to a preset rule, and selectively inserting preset conjunctions or separators between each dimension information; The preset rules include at least the importance of information, keyword guidance, natural language connection and structured tags.

7. A system for AI instruction structured design method based on the 5W3H framework according to any one of claims 1 to 6, characterized in that: The system comprises: An information receiving module is used to obtain user demand information; the user demand information is preliminary information input based on the 5W3H framework; An instruction combination module, configured to combine the preliminary information of each dimension input by the user into a structured AI instruction according to predetermined rules; wherein the AI ​​instruction also includes an emotional tone adapted to a specific application scenario; The instruction output module is used to output the structured AI instructions.

8. The system according to claim 7, characterized in that The system further comprises: The element enhancement module is arranged between the information receiving module and the instruction combination module, and is used to call at least one preset artificial intelligence model, perform multi-dimensional processing on the preliminary information, generate element content, and pass the element content to the instruction combination module to optimize the user demand information.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, it implements the AI ​​instruction structured design method based on the 5W3H framework as described in any one of claims 1 to 6.

10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, it implements the AI ​​instruction structured design method based on the 5W3H framework as described in any one of claims 1-6.

Images