The development of artificial intelligence models like ChatGPT, Gemini, and Grok has transformed the way people approach problem-solving, seek inspiration, and create content, whether text, images, or videos. However, many users are unaware that the quality of the responses provided by these platforms is intrinsically linked to the formulation of the instructions given to them, known as prompts.
What is a prompt?
A prompt is defined as the written instruction provided to a chatbot. These instructions can vary from simple requests, such as generating a recipe, to more complex requirements that include context, audience, style, or format. Improving the writing of prompts can revolutionize the use of artificial intelligence tools.
Developing effective techniques for creating prompts is crucial for ensuring that the interaction between humans and artificial intelligence systems is effective, efficient, and personalized. Experts in the field, such as content generator Franco Dumont, suggest that following structured frameworks can help craft better instructions and optimize the results obtained.
Methodologies for Creating Effective Prompts
ROCKET Method
The ROCKET method provides a set of steps that help users not overlook essential elements in the instruction directed at an artificial intelligence platform. This English acronym breaks down each key aspect of the prompt:
- R (Role): Clearly defining the role that the AI will assume allows for receiving responses from a specific perspective. For example: "You are a corporate communication specialist."
- O (Objective): Precisely specifying the required task, such as drafting a newsletter or explaining a technical concept.
- C (Context): Including relevant information that the AI may not have by default, such as company data or the type of audience.
- K (Knowledge Limits): Establishing the limits on what is not desired in the response, known as a "negative prompt," helps avoid jargon and sensitive topics.
- E (Expression Style): Indicating the tone, register, or style expected in the response.
- T (Type of Output): Defining the expected format of the response, whether it be a list, a document, or an infographic.
This structure transforms a general request into clear instructions that the artificial intelligence can follow without deviations.
ROCEF Technique
The ROCEF technique, which predates ROCKET, shares the premise that the more detailed each component of the request is, the better the results will be. This method consists of five fundamental elements:
- R (Role): Indicating the role that the AI will play in the response.
- O (Objective): Clarifying the outcome that is desired.
- C (Context): Incorporating relevant information about the environment or purpose of the task.
- E (Examples): Providing models that the AI can use as a reference, whether a text or an image.
- F (Format): Specifying the expectation for the outcome, such as a short text or a visual summary.
Prompt Examples Using ROCKET and ROCEF
Prompt with the ROCKET Methodology
Example: "You are a human resources specialist with ten years of experience in recruiting for technology companies. I need you to write a practical guide for attracting and retaining young talent in the IT sector. The company I work for is located in Madrid, specializes in mobile application development, and has a workforce of fifty employees. Do not include generic advice or use complex technical jargon. Use a professional and straightforward tone, focused on clarity and impact. The final result should be a list of five detailed recommendations."
ROCKET Explanation:
- R (Role): The AI acts as an HR expert for the technology sector.
- O (Objective): A guide to attract and retain young talent is requested.
- C (Context): The location, type of company, and size of the team are specified.
- K (Knowledge Limits): Generic advice and technical jargon are not allowed.
- E (Expression Style): A professional and clear tone is sought.
- T (Type of Output): A numbered list of five recommendations is required.
Prompt with the ROCEF Methodology
Example: "Act as a digital marketing consultant specializing in SMEs. The objective is to design a campaign to promote a sustainable clothing e-commerce in Barcelona that aims to expand its customer base among young people. I provide you with an example of a previous campaign: ‘Eco discounts on fashion for university students during September, with plastic-free shipping and part of the sale donated to local NGOs.’ The result I need is a creative proposal in the format of an advertising brief of no more than 300 words."
ROCEF Explanation:
- R (Role): The AI responds as a digital marketing consultant for SMEs.
- O (Objective): Develop a campaign to attract young customers for a sustainable clothing e-commerce.
- C (Context): The location (Barcelona) and the focus on sustainability are mentioned.
- E (Examples): A previous campaign is included as a reference.
- F (Format): A proposal is requested in the format of an advertising brief of 300 words.
Conclusions
The advancement of chatbots demands that users and professionals learn to articulate their requests with greater detail and precision, using structured frameworks like ROCKET and ROCEF. This not only enhances the quality of the responses generated by artificial intelligences but also optimizes the interaction experience with these technologies.
To read more about the application of artificial intelligence and its impact in various areas, visit the personal blog.
