19 Formulas and Prompt Structures for ChatGPT: Going Beyond the Basics

In the ongoing journey to enhance communication between humans and artificial intelligences, like ChatGPT, one tool stands out for its ability to shape rich and productive interactions: prompt engineering. Beyond the basics previously explored, a particularly fascinating aspect of this practice involves the use of formulas and prompt structures, true catalysts for maximizing the effectiveness and accuracy of responses generated by AI.

This article is dedicated exclusively to delving into this topic, unveiling a range of structural strategies from basic approaches to the most advanced ones. By understanding and applying these formulas, it’s possible not just to communicate more efficiently with ChatGPT and other AI tools, such as Microsoft Copilot and Google Gemini, but also to explore new dimensions of creativity, analysis, and problem-solving.

Each prompt structure offers a unique way of organizing and presenting your requests, guiding the AI model to generate responses that closely align with your specific objectives. Whether to elicit detailed information, foster creative content generation, or facilitate the resolution of complex issues, these strategies represent a bridge between human intent and the nearly limitless potential of artificial intelligence.

Throughout this article, we’ll explore various formulas and structures, highlighting their distinctive features, practical applications, and, most importantly, how they can be combined and adapted to meet a wide range of needs and contexts. Prepare to dive into the universe of prompt formulas and structures, a journey that promises not only to expand your prompt engineering repertoire but also to enrich your ability to interact more meaningfully with ChatGPT.

Free eBook: Prompt Engineering Unveiled

Master the art of prompts in 43 pages and unlock the potential of LLM AIs.

Download Now

Examples of Formulas and Prompt Structures

RTF: Role, Task, Format

The RTF structure is an efficient approach for formulating prompts that direct the interaction with ChatGPT in a clear and objective manner. This methodology helps establish the function or role that ChatGPT should assume (Role), the specific activity or problem to be solved (Task), and how the response should be organized (Format).

Utilizing the RTF structure, it’s possible to guide ChatGPT’s responses to be more in line with user expectations, besides facilitating the acquisition of information or solutions structured in a particularly useful way. Let’s explore each of its components:

• Role: Defines who or what is performing the action in the prompt. It can be a person, an entity, or ChatGPT itself acting in a specific role.
• Task: Specifies what needs to be done. The task should be clear and direct, describing the desired action or necessary information.
• Format: Outlines how you wish the response to be structured. This can include formats such as a numbered list, a paragraph, a mathematical formula, etc.

Practical example of RTF:

As a nutrition expert (Role), provide a list of five protein-rich foods (Task) in a numbered list format (Format).

ChatGPT might respond something like:

1. Chicken
2. Eggs
3. Quinoa
4. Almonds
5. Greek Yogurt

This example demonstrates how the RTF structure can be used to generate a specific and well-organized response, facilitating the communication of clear and concise information. By specifying the role (nutrition expert), the task (provide a list of foods), and the format (numbered list).

CTF: Context, Task, Format

The CTF structure is a powerful tool in prompt engineering, allowing users to clearly establish the scenario in which the interaction occurs, the specific action expected from ChatGPT, and how the response should be structured. This approach facilitates the creation of prompts that are both precise in their requests and specific regarding the desired outcome, optimizing the effectiveness of the response obtained.

• Context: Provides the necessary background information to understand the situation or problem at hand. Helps establish the relevance of the task and guides the direction of the response.
• Task: Explicitly describes what ChatGPT is expected to do, detailing the action or set of actions required.
• Format: Specifies how the response should be structured, whether in the form of a list, paragraphs, a detailed plan, etc., to meet the user’s needs.

Suppose you’re seeking help from ChatGPT to compile a list of practical energy-saving tips for homes. Applying the CTF structure, the prompt would be:

• Context: “Given the increase in energy tariffs and growing concerns about environmental sustainability, many people are looking for ways to reduce energy consumption in their homes.”
• Task: “Compile a list of practical and easy-to-implement tips that homeowners can use to save energy.”
• Format: “Present the tips in a numbered list format, providing a brief description for each.”

Practical example of CTF:

Considering the increase in energy tariffs and concern for sustainability, create a numbered list of practical energy-saving tips for homes. The tips should be easy to implement and accompanied by a brief explanatory description.

This example shows how the CTF structure can be used to request specific information in a clear and organized manner, resulting in a response that is directly applicable and useful for the presented context.

PECRA: Purpose, Expectation, Context, Request, Action

The PECRA structure is a versatile tool for prompt engineering that emphasizes clarity and specificity when interacting with language models like ChatGPT. Let’s break down each component:

• Purpose: Defines the reason for creating the prompt. It helps clarify the overall intention of the interaction.
• Expectation: Describes the outcome or type of response expected from the model.
• Context: Provides additional information necessary for the model to understand the prompt and generate an appropriate response.
• Request: Clearly specifies what is being requested from the model.
• Action: Indicates the specific action you want the model to take.

Imagine you want ChatGPT to create a study plan for a student preparing for an important math exam. Here’s how to apply the PECRA structure:

• Purpose: “The purpose of this prompt is to help a student effectively prepare for a math exam.”
• Expectation: “I expect to receive a detailed study plan covering the main topics needed for the exam.”
• Context: “The student has 2 weeks until the exam, can study approximately 3 hours per day, and struggles mainly with algebra and geometry.”
• Request: “Based on this information, please create a study plan.”
• Action: “Organize the plan starting with algebra fundamentals, moving on to geometry, and include regular reviews.”

Practical example of PECRA:

Considering a student who is preparing for an important math exam in 2 weeks and can dedicate 3 hours daily to studying, with difficulties in algebra and geometry, create a detailed study plan. The plan should start with the fundamentals of algebra, progress to geometry, and include regular reviews, aiming for effective preparation for the exam.

This example illustrates how the PECRA structure can be used to formulate a clear and detailed prompt, guiding ChatGPT to generate a response that meets the user’s specific expectations.

CREATE: Character, Request, Examples, Adjustments, Type, Extras

The CREATE structure is a comprehensive method for formulating prompts, aiming for efficient and targeted interaction with AI models like ChatGPT. This detailed approach allows the user to specify the context and expectations of the interaction clearly. Let’s look at each component:

• Character: Defines a specific role for ChatGPT, guiding its responses according to a predetermined profile or function. This helps shape the nature of the responses according to the desired context.
• Request: Specifies in a clear and objective manner what is expected from the model, detailing the task to be performed. This step is crucial to ensure the model understands exactly what is requested.
• Examples: Provides examples of outputs or expected results, giving the model a clear reference of the type of response or content desired.
• Adjustments: Allows the user to request specific improvements or modifications in previous responses, directing the model to better meet the needs of the task.
• Type of output: Defines the expected format of the response, whether it’s narrative text, a list, a detailed plan, among others. This guides how the information provided by the model is structured.
• Extras: Offers the opportunity to add additional or contextual information, enriching the prompt and enabling more precise and aligned responses with the user’s expectations.

Imagine you want ChatGPT to create a personalized travel guide for a city you plan to visit. Here’s how to apply the CREATE structure:

• Characterization: “Acting as an experienced local travel guide,”
• Request: “create a personalized guide.”
• Examples: “Include categories like accommodations, gastronomy, tourist attractions, and transportation tips.”
• Adjustments: “Prioritize options that are budget-friendly and suitable for families.”
• Type: “Organize the guide into clearly defined sections, with recommendations and brief descriptions.”
• Extras: “I’m traveling in July, so include relevant seasonal events and activities.”

Practical example of CREATE:

Acting as an experienced local travel guide, create a personalized guide for my trip to Barcelona in July. Include accommodations, gastronomy, tourist attractions, and transportation tips, prioritizing options that are budget-friendly and suitable for families. Organize the guide into clearly defined sections, with recommendations and brief descriptions for each item, and don't forget to add relevant seasonal events and activities for the period of my visit.

This example demonstrates how to use the CREATE structure to elaborate a detailed and specific prompt, guiding ChatGPT to produce a response aligned with the user’s expectations and needs.

CREO: Context, Request, Explanation, Outcome

The CREO structure is a methodology focused on optimizing prompt formulation for more effective interactions with language models, like ChatGPT. It emphasizes the importance of context, a clear request, an explanation of the task, and the definition of the expected outcome.

• Context: Provides the background information necessary for the model to understand the situation or topic addressed.
• Request: Clearly specifies what is expected from the model.
• Explanation: Explains the task in detail, helping the model better understand the purpose of the request.
• Outcome: Describes the type of response or result expected from the model.

Imagine you want ChatGPT to create a list of suggestions to increase personal productivity. Here’s how to apply the CREO structure:

• Context: “Considering many people work from home and face frequent distractions,”
• Request: “create a list of suggestions.”
• Explanation: “These suggestions should be practical and easy to implement for those working from home.”
• Outcome: “I expect a list that includes time management techniques, workspace setup, and wellness tips.”

Practical example of CREO:

Considering many people work from home and face frequent distractions, create a list of practical and easy-to-implement suggestions to increase personal productivity. These suggestions should cover time management techniques, workspace setup, and wellness tips, aiming to improve focus and efficiency in the home environment.

This example illustrates how the CREO structure can be utilized to formulate clear and objective prompts, guiding ChatGPT to produce responses that effectively meet the user’s specific needs.

PAIN: Problem, Action, Information, Next Steps

The PAIN structure is a prompt engineering methodology that focuses on identifying and solving specific problems through ChatGPT. It guides the formulation of requests to extract precise and applicable solutions.

• Problem: Identifies the problem that needs to be solved, clarifying the challenge or need of the user.
• Action: Specifies the action or type of help expected from ChatGPT, directing it towards solving the problem.
• Information: Requests detailed information or clarifications that ChatGPT can provide to better understand the context or nuances of the problem.
• Next Steps: Asks for an action plan, resources, or subsequent steps the user can follow to solve the problem or achieve the desired goal.

Imagine you’re struggling to efficiently organize your time and want ChatGPT to help create a time management plan. Here’s how to apply the PAIN structure:

• Problem: “I’m struggling to manage my time effectively,”
• Action: “I need a personalized time management plan.”
• Information: “What strategies or tools would you recommend?”
• Next Steps: “Provide a step-by-step plan that I can start following immediately.”

Practical example of PAIN:

I'm struggling to manage my time effectively and need a personalized time management plan. What strategies or tools would you recommend? Please provide a step-by-step plan that I can start following immediately, considering my day is often interrupted by unexpected tasks.

This example shows how the PAIN structure can be applied to create a directed and effective prompt, leading ChatGPT to offer practical and personalized solutions for specific problems faced by the user.

TREF: Task, Requirement, Expectation, Format

The TREF structure is a focused approach to prompt engineering that helps specify what you want to achieve with the interaction, the criteria that must be met, what you expect from the response, and in what format it should be delivered. Let’s explore each of its components:

• Task: What exactly you’re asking ChatGPT to do. This should be a clear action or set of actions.
• Requirement: The specific criteria or conditions the response needs to satisfy.
• Expectation: The expected outcome of the interaction, including the type of information or solution you want to receive.
• Format: How you want the information or solution to be presented.

Suppose you want ChatGPT to write a summary about the current trends in renewable energy technology. Here’s how to apply the TREF structure:

• Task: “Write a summary about the current trends in renewable energy technology.”
• Requirement: “The summary must cover both technological advancements and current challenges.”
• Expectation: “I expect to get a clear and concise overview that can be used to inform the general public.”
• Format: “The summary should be structured in paragraphs, with no more than 300 words.”

Practical example of TREF:

Please write a summary of no more than 300 words about the current trends in renewable energy technology, including technological advancements and challenges. The summary should be structured in paragraphs, offering a clear and concise overview suitable for informing the general public.

This example shows how the TREF structure can be used to create a detailed and specific prompt, directing ChatGPT to produce a response that not only meets a specific informational need but also fulfills defined content and formatting criteria.

GRADE: Goal, Request, Action, Detail, Examples

The GRADE structure is an effective technique for structuring prompts that clearly communicate the intention, detail the request, specify the desired action, provide additional details, and include examples to guide the response. Here is a breakdown of each component:

• Goal: The objective or purpose of the prompt. Defines what you hope to achieve with the interaction.
• Request: What you’re specifically asking ChatGPT to do.
• Action: The specific action you want ChatGPT to take in response to your request.
• Detail: Additional information that helps clarify the request, providing more precise context or specifications.
• Examples: Cases or concrete examples that illustrate the type of response or content you expect to receive.

Let’s imagine you want ChatGPT to create an introductory guide for beginners on how to invest in cryptocurrencies. Here’s how the GRADE structure can be applied:

• Goal: “Create an accessible introductory guide for beginners on cryptocurrency investment.”
• Request: “Develop a guide that introduces the basic concepts of cryptocurrency investment.”
• Action: “Include sections on what cryptocurrencies are, how to start investing, and security tips.”
• Detail: “The guide should be easy to understand for someone with no prior knowledge of the subject.”
• Examples: “Provide examples of popular investment platforms and explain common terms like ‘blockchain’ and ‘digital wallet’.”

Practical example of GRADE:

Please create an introductory guide on how to invest in cryptocurrencies, aimed at beginners. The guide should introduce basic concepts, include sections on what cryptocurrencies are, how to start investing, and security tips. Ensure the content is accessible to someone with no prior knowledge, providing examples of popular investment platforms and explaining terms like 'blockchain' and 'digital wallet'.

This example demonstrates how the GRADE structure can be utilized to elaborate a detailed and specific prompt, guiding ChatGPT to produce a comprehensive and accessible introductory guide on a complex topic like cryptocurrency investment.

ROSES: Role, Objective, Scenario, Expected Solution, Steps

The ROSES structure is designed to facilitate detailed communication of a problem and how you would like it to be approached, specifying the role, objective, scenario in which the issue is placed, the expected solution, and the steps to achieve it. This approach is particularly useful for complex requests or when a detailed and structured response is desired. Let’s detail each component:

• Role: Defines who is performing the action or from whose perspective. It can be ChatGPT assuming a specific role.
• Objective: What is expected to be achieved with the prompt. Clarifies the purpose of the request.
• Scenario: The context or situation in which the question or task is placed. Helps provide foundation and relevance to the request.
• Expected Solution: Describes the outcome or type of response expected from ChatGPT.
• Steps: A sequence of actions or process that should be followed to reach the desired solution.

Imagine you want ChatGPT to help plan a digital marketing campaign for a new product. Using the ROSES structure, the prompt could be structured as follows:

• Role: “As a digital marketing expert…”
• Objective: “…the goal is to create an effective campaign for the launch of a new technological product.”
• Scenario: “The product is an innovative health monitoring device that connects to smartphones. The target market is young adults interested in technology and fitness.”
• Expected Solution: “An expected campaign plan that includes social media strategies, digital influencers, and email marketing.”
• Steps: “1. Identify the main social media platforms used by our target audience. 2. Select digital influencers in the technology and fitness niche. 3. Develop a series of emails for pre and post-launch engagement.”

Practical example of ROSES:

As a digital marketing expert, create a digital marketing campaign for the launch of a new health monitoring device that connects to smartphones. The product is aimed at young adults interested in technology and fitness. The plan should include strategies for social media, digital influencers, and email marketing, starting by identifying the main social media platforms, selecting influencers in the technology and fitness niche, and developing a series of emails for engagement before and after the launch.

RDIREC: Role, Definition, Intent, Request, Example, Clarification

The RDIREC structure is a detailed methodology for formulating prompts that require complex and well-founded responses, incorporating elements such as the role played, the definition of key terms, the intention behind the request, specific examples to guide the response, and clarifications to avoid ambiguities. Let’s explore each component:

• Role: Specifies the viewpoint or capacity in which ChatGPT or the user is acting.
• Definition: Clarifies key concepts or terms that are crucial for understanding the prompt.
• Intent: Explains the reason or goal behind the prompt, which helps guide the direction of the response.
• Request: What exactly is being requested, formulated in a clear and precise manner.
• Example: Provides cases or concrete examples that serve as a reference for the type of response expected.
• Clarification: Adds details or additional information to minimize misunderstandings and refine the response.

Suppose you want ChatGPT to create content about the importance of cybersecurity for small and medium enterprises (SMEs). Applying the RDIREC structure, the prompt can be structured as follows:

• Role: “As a cybersecurity consultant…”
• Definition: “…define ‘cybersecurity’ and explain its relevance to the current business environment, especially for SMEs.”
• Intent: “The goal is to raise awareness among SME owners about cyber risks and encourage them to adopt protective measures.”
• Request: “Develop an introductory guide on cybersecurity for SMEs, highlighting best practices and risk mitigation strategies.”
• Example: “Include examples of common cyberattacks, such as phishing and ransomware, and their consequences for businesses.”
• Clarification: “Emphasize the importance of employee training, regular backups, and software updates as preventive measures.”

Practical example of RDIREC:

As a cybersecurity consultant, define 'cybersecurity' and explain its importance in the current business context, focusing on SMEs. The goal is to create an introductory guide that raises awareness about cyber risks and promotes the adoption of security measures. Include examples of attacks like phishing and ransomware, highlighting the consequences for businesses. Detail the relevance of employee training, regular backups, and software updates as preventive strategies.

This example shows how the RDIREC structure can be used to formulate a complex prompt, guiding ChatGPT to produce educational and detailed content on cybersecurity for SMEs, with clear definitions, illustrative examples, and clarifications that effectively direct the response.

RSCET: Role, Situation, Complication, Expectation, Task

The RSCET structure is used to develop prompts that detail a complex scenario, requiring a response that addresses a specific situation, its complications, what is expected as a solution, and the task to be performed. This methodology helps to create clear and structured prompts for scenarios involving problem-solving or detailed analysis. Let’s detail each component:

• Role: Defines who is involved or who should act, which could be ChatGPT assuming a specific role.
• Situation: Describes the context or scenario in which the prompt is placed.
• Complication: Identifies the challenges, problems, or complications present in the described situation.
• Expectation: Clarifies what is expected as a result of the interaction, what solution or type of response is desired.
• Task: Specifies the action or set of actions that must be performed to address the prompt.

Imagine you want ChatGPT to help plan a strategy to overcome a digital marketing challenge faced by a technology startup. Using the RSCET structure, the prompt could be formulated as follows:

• Role: “As a digital marketing expert…”
• Situation: “…the technology startup ‘TechNova’ is launching a new productivity app, which helps users better manage their time and projects.”
• Complication: “Despite the high quality of the app, ‘TechNova’ faces strong competition in the market and has difficulties reaching its target audience.”
• Expectation: “An innovative digital marketing strategy is expected to highlight the app in the saturated market and increase its reach.”
• Task: “Create a marketing plan that includes SEO tactics, content marketing, and social media campaigns, focusing on the app’s differentiators and how it solves specific user problems.”

Practical example of RSCET:

As a digital marketing expert, develop a strategy for the startup 'TechNova,' which is launching a new productivity app. Despite the quality of the product, the company faces strong competition and challenges in reaching its audience. The goal is to create a marketing plan that utilizes SEO, content marketing, and social media, highlighting the app's differentiators and its ability to solve user problems.

This example demonstrates how the RSCET structure can be applied to elaborate a prompt that details a challenging scenario, directing ChatGPT to develop a creative and focused digital marketing strategy to overcome the specified complications.

RASCEF: Role, Action, Steps, Context, Examples, Format

The RASCEF structure is a detailed approach for formulating prompts that emphasize clarity in communicating a complex task, incorporating the role assumed, the actions to be taken, the specific steps for accomplishing the task, the context surrounding the situation, examples to better illustrate the request, and the desired format for the response. This structure is ideal for situations requiring detailed instructions and well-defined results. Let’s examine each component:

• Role: Defines who is performing the task or from whose perspective the situation is described.
• Action: Describes the main action or actions that need to be taken.
• Steps: Details the specific steps or procedures needed to complete the task.
• Context: Provides relevant background information to fully understand the situation or problem at hand.
• Examples: Includes examples or practical cases that serve as a model or inspiration for the response.
• Format: Specifies how the response should be organized or presented.

Imagine you want ChatGPT to develop a plan to increase the online visibility of a new artisan coffee brand. Using the RASCEF structure, the prompt could be structured as follows:

• Role: “As a digital marketing consultant specialized in artisan coffee brands…”
• Action: “…develop a strategic plan to increase the brand’s online visibility.”
• Steps: “1. Identify the target audience. 2. Choose the most suitable social media platforms. 3. Create engaging content that highlights the uniqueness of the coffee. 4. Implement a paid advertising campaign. 5. Measure and adjust the strategy based on feedback and data analysis.”
• Context: “The brand is new to the market and offers a unique selection of single-origin artisan coffees but is struggling to stand out in a competitive market.”
• Examples: “Include examples of content types that can be created, such as posts about the coffee’s origin, behind-the-scenes videos showing the roasting process, and customer testimonials.”
• Format: “The plan should be presented in a structured document with clear sections for each step of the process.”

Practical example of RASCEF:

As a digital marketing consultant specialized in artisan coffee brands, devise a strategic plan to increase the online visibility of a new coffee brand. The plan should include steps to identify the target audience, select suitable social media platforms, create engaging content, implement paid ad campaigns, and measure the success of the strategy. Consider the context of a competitive market and provide examples of content. Present the plan in a structured document with sections for each step.

This example demonstrates how the RASCEF structure facilitates the creation of a detailed and action-oriented prompt, allowing ChatGPT to generate a comprehensive and well-organized digital marketing plan for the artisan coffee brand.

APE: Action, Purpose, Expectation

The APE structure is a concise methodology focusing on three crucial elements for formulating effective prompts: the desired action, the purpose behind this action, and the outcome expectation. Let’s detail each component:

• Action: What you want ChatGPT to do. This component is straightforward and clearly specifies the task to be performed.
• Purpose: The reason you are requesting this action. It defines the intention behind the prompt, clarifying the goal you seek to achieve.
• Expectation: The outcome you expect to see as a response to the requested action. It specifies what would be considered a satisfactory response.

Imagine you want ChatGPT to create informative content about the impact of artificial intelligence on education. Using the APE structure, the prompt would be:

• Action: “Write an article on the impact of artificial intelligence on education.”
• Purpose: “The purpose is to inform readers about how AI is transforming teaching and learning methods.”
• Expectation: “I expect a detailed analysis covering both the benefits and challenges associated with the use of AI in education, with concrete examples.”

Practical example of APE:

Write a detailed article about the impact of artificial intelligence on education, aiming to inform readers about the transformations in teaching and learning methods. I expect an analysis that explores the benefits and challenges, including concrete examples to illustrate these points.

This example shows how the APE structure can be utilized to create a clear and objective prompt, directing ChatGPT to produce informative and well-founded content on the proposed topic.

TAG: Task, Action, Goal

The TAG structure is an efficient tool for defining prompts that are focused and outcome-oriented. It emphasizes the importance of establishing a specific task, the necessary action to complete it, and the final goal to be achieved. This approach helps to create a clear direction for the interaction with ChatGPT. Let’s detail each component:

• Task: What needs to be done. This element defines the scope of the work or the information requested.
• Action: The specific steps or process by which the task should be accomplished.
• Goal: The desired outcome or the purpose of completing the task. It clarifies what is expected to be achieved at the end of the process.

Suppose you are seeking guidance from ChatGPT to plan a networking event for technology professionals. Using the TAG structure, the prompt could be:

• Task: “Organize a networking event for technology professionals.”
• Action: “Identify the key elements necessary for the event’s success, such as venue, discussion themes, and special guests.”
• Goal: “Create a networking opportunity that fosters valuable exchanges among participants and fosters lasting professional connections.”

Practical example of TAG:

Organize a networking event for technology professionals. To do this, identify the key elements that will contribute to the event's success, including the choice of venue, relevant discussion themes, and the selection of special guests. The objective is to create an environment that promotes valuable exchanges and fosters lasting professional connections among participants.

This example illustrates how the TAG structure can be applied to generate a clear and objective plan for organizing an event, detailing the task, necessary actions, and the desired final goal.

ERA: Expectation, Role, Action

The ERA structure focuses on clearly defining the expectation for the response, the role that ChatGPT or the user plays in the interaction, and the specific actions that must be taken to meet the expectation. This approach effectively guides the language model, ensuring that responses are aligned with the user’s goals. Let’s explore each component:

• Expectation: The desired outcome or what is expected to be obtained as a response. This element establishes the interaction’s final goal.
• Role: The function or identity assumed by ChatGPT or the user in the context of the prompt. Specifying the role helps to contextualize the response within a particular scenario or perspective.
• Action: The steps or processes that must be followed to fulfill the task and meet the established expectation.

Imagine you want ChatGPT to help you develop a study plan for an information technology certification exam. Applying the ERA structure, the prompt would be:

• Expectation: “Develop an effective study plan that covers all the necessary topics for the information technology certification exam within a three-month period.”
• Role: “As a virtual tutor experienced in preparing for information technology certification exams…”
• Action: “…create a detailed study schedule, including recommended learning resources, a balanced distribution of topics over the period, and effective revision techniques.”

Practical example of ERA:

As a virtual tutor specializing in preparing for information technology certification exams, develop a detailed study plan for a certification exam taking place in three months. The plan should include a study schedule, recommended learning resources, a balanced distribution of topics, and effective revision techniques, aiming to cover all necessary topics for the exam.

This example demonstrates how the ERA structure can be utilized to request ChatGPT to create a detailed and structured study plan, clearly establishing the expectation, the role played by the model, and the specific actions to be taken.

RACE: Role, Action, Context, Expectation

The RACE structure is a comprehensive methodology for creating prompts that emphasize defining the role assumed in the interaction, the desired action, the context surrounding the request, and the expectation of the outcome. This approach ensures that all essential parts of a prompt are considered, facilitating accurate and aligned responses with the user’s objectives. Let’s detail each component:

• Role: Specifies who is performing the action or the perspective from which the request is made. It can be ChatGPT assuming a specific role or the user defining their position or function.
• Action: Clearly describes what is expected of ChatGPT to do, detailing the task or actions required.
• Context: Provides background information and additional details about the situation or problem at hand, helping to clarify the prompt and guide the response.
• Expectation: Defines the desired outcome or what is expected to be achieved with the response, establishing a clear goal for the interaction.

Imagine you are seeking assistance from ChatGPT to optimize the layout of an e-commerce website to improve user experience. Using the RACE structure, the prompt could be formulated as follows:

• Role: “As a UX (User Experience) designer…”
• Action: “…evaluate the current layout of our e-commerce website and suggest specific improvements.”
• Context: “The site has a high cart abandonment rate, and user feedback indicates that finding products can be challenging.”
• Expectation: “I expect to receive concrete suggestions that can be implemented to simplify navigation, make product search more intuitive, and consequently reduce the cart abandonment rate.”

Practical example of RACE:

As a UX designer, evaluate the current layout of our e-commerce website, considering that we face a high cart abandonment rate and user feedback indicating difficulties in finding products. Based on this, suggest specific improvements that can make navigation simpler and product search more intuitive, with the goal of reducing the cart abandonment rate.

This example shows how the RACE structure can be utilized to formulate a detailed and directed prompt, requesting ChatGPT for a critical analysis and suggestions for improvements for an e-commerce website, based on a specific context and with a clear expectation of outcome.

COAST: Context, Objective, Actions, Scenario, Task

The COAST structure is designed to guide the formulation of prompts in a way that covers all the essential aspects of a complex request, from the context to the specific task to be performed. This methodology helps to ensure that instructions are clear and complete, facilitating the generation of precise and relevant responses by ChatGPT. Let’s explore each of its components:

• Context: Provides the necessary background information to understand the situation or problem at hand, establishing the foundation for the interaction.
• Objective: Clearly defines what is expected to be achieved with the interaction, clarifying the purpose of the request.
• Actions: Details the steps or specific processes that must be followed to fulfill the objective.
• Scenario: Describes the specific situation or set of circumstances in which the task is set, providing additional context.
• Task: Specifies the action or set of concrete actions that ChatGPT must perform to address the prompt.

Suppose you want ChatGPT to help develop a strategy to increase employee participation in a corporate wellness program. Applying the COAST structure, the prompt might be:

• Context: “Our company recently launched a wellness program to promote physical and mental health among employees, but participation has been lower than expected.”
• Objective: “The goal is to develop an effective strategy to increase employee participation in the wellness program.”
• Actions: “Identify barriers to participation, create incentives for participation, and develop effective communication channels to inform and engage employees.”
• Scenario: “Considering that many employees work remotely and may not be aware of all the benefits of the program.”
• Task: “Devise a detailed plan that includes specific actions to overcome the identified barriers, incentives for participation, and communication strategies to increase awareness about the program.”

Practical example of COAST:

Given the context of low participation in the recently launched corporate wellness program, develop a strategy to increase employee engagement. Identify the main barriers to participation, propose attractive incentives, and structure effective communication channels, considering the scenario of many employees working remotely. The task is to create a detailed plan that addresses these points, aiming to significantly improve participation in the program.

This example demonstrates how the COAST structure can be used to construct a detailed and comprehensive prompt, guiding ChatGPT to develop a complex strategy involving multiple aspects, from problem analysis to the proposition of concrete solutions.

RISE: Role, Input, Steps, Expectation

The RISE structure is designed to guide the formulation of prompts that require a step-by-step approach to achieving a desired outcome, focusing on the interlocutor’s role, the necessary inputs to start the action, the detailed steps of the process, and the expectation of the result. This methodology is particularly useful for tasks involving complex instructions or processes. Let’s detail each component:

• Role: Defines the function or perspective of whoever is performing the task, be it the user or ChatGPT.
• Input: Specifies the information or resources required to initiate the task or process.
• Steps: Details the consecutive actions that must be taken to complete the task or achieve the goal.
• Expectation: Clarifies the outcome or goal expected to be reached at the end of the process.

Suppose you want ChatGPT to help plan and execute a market research for a new product. Using the RISE structure, the prompt could be:

• Role: “As a market research analyst…”
• Input: “…with access to demographic data of the target audience and online survey tools…”
• Steps: “1. Clearly define the target audience for the survey. 2. Draft a questionnaire focusing on key aspects of the new product. 3. Choose the most suitable online survey platform. 4. Analyze the collected data to extract relevant insights.”
• Expectation: “I expect to receive a detailed report with insights on the product’s acceptance in the target market, including marketing strategy recommendations.”

Practical example of RISE:

As a market research analyst, with access to demographic data of the target audience and online survey tools, plan and execute a survey for a new product. The process should include defining the target audience, drafting a questionnaire, choosing the survey platform, and analyzing the collected data. The goal is to obtain a report with insights on the product's acceptance in the market, accompanied by recommendations for the marketing strategy.

This example demonstrates how the RISE structure can be effectively applied to create a detailed and action-oriented prompt, facilitating the completion of a complex task, such as planning and executing market research, by following specific steps to achieve a well-defined outcome.

SPARK: Situation, Problem, Aspiration, Results, Kismet

The SPARK structure is a rich, narrative approach to prompt formulation, designed to deeply explore a situation, identify problems, define aspirations, anticipate outcomes, and consider the element of luck or destiny (kismet). This methodology is ideal for complex scenarios where a full understanding of the context and goals is crucial for generating creative and effective solutions. Let’s detail each component:

• Situation: Describes the current context or scenario that serves as the backdrop for the interaction or the issue at hand.
• Problem: Identifies the central challenge or issue that needs to be addressed or resolved.
• Aspiration: Defines the ideal state or goal that is desired to be achieved, contrasting with the current problem.
• Results: Anticipates the positive outcomes or achievements expected from solving the problem or reaching the aspiration.
• Kismet: Considers the element of luck, destiny, or factors outside of control that might influence the outcome of the situation.

Imagine you’re seeking ChatGPT’s help to develop an innovative strategy to increase engagement with a wellness app. Applying the SPARK structure, the prompt could be:

• Situation: “The wellness app ‘WellLife’ has been receiving positive feedback on its interface and functionalities, but user engagement has been low in recent months.”
• Problem: “The main challenge is the lack of ongoing engagement from users, who often download the app but don’t use it regularly.”
• Aspiration: “Our goal is to make ‘WellLife’ an essential part of users’ daily wellness routine, significantly increasing engagement and participation.”
• Results: “We expect to see an increase in the app’s usage frequency, better user retention, and positive feedback on the new strategies implemented.”
• Kismet: “We acknowledge that external factors, such as market trends and new technologies, could influence the success of our initiatives.”

Practical example of SPARK:

Given the current situation of the wellness app 'WellLife', with positive feedback but low user participation, the challenge is to increase continuous engagement. We aspire to make 'WellLife' an indispensable part of users' wellness routines. We hope, as a result, to see an increase in daily use and better user retention, although we recognize that external factors may impact these outcomes. Develop an innovative strategy to achieve these goals.

This example demonstrates how the SPARK structure can be utilized to craft a detailed, solution-oriented prompt, addressing a complex situation with a mix of deep analysis, clear objectives, anticipation of positive outcomes, and a notion of the role of fate or luck.

Frequently Asked Questions

Conclusion

As we explored the vast territory of formulas and prompt structures in this article, we unveiled essential tools that empower users to communicate more effectively with advanced language models like ChatGPT. The discussed formulas offer a range of options for structuring requests in a way that maximizes the accuracy, relevance, and depth of the obtained responses.

Prompt engineering is not just about asking questions; it’s about building bridges of understanding between humans and artificial intelligence. By applying the covered techniques, from defining contexts and expectations to detailing steps and clear goals, we pave the way for richer and more productive collaboration with AI. Each presented structure serves as a guide, a mold that can be adapted and personalized according to the unique needs of each interaction, promoting a more aligned and efficient exchange of information.

We encourage readers not only to apply these structures in their interactions with ChatGPT but also to experiment with their combinations and variations. Flexibility and adaptability are central features of prompt engineering, and careful experimentation may reveal innovative approaches and creative solutions to communication challenges.