Director, Learning Design & Development| PMIAsheville, NC, United States
With Generative AI, iteratively refining and optimizing prompts can lead to better AI-generated results. This may involve adjusting the specificity or clarity of the prompt to increase relevance and accuracy of results.
What examples do you have of how improving a prompt drastically changed the output quality? What specific changes did you make that led to the improvement?
The more specific the prompt, the better the results. For instance, if your prompt is "workout tips" the response will be very generic and not tailored to your needs. Refining it to "strength-training workouts for women over 40 will give a must more robust answer with actionable items. Saving Changes...
In my experience, refining "persona input" and attached "output format" example matter for GenAI output. Saving Changes...
Valerie BrownProgram Management| Anthem Inc.Fort Myers, FL, United States
In my experience, refining a prompt is crucial in obtaining high-quality outputs from GenAI systems. It involves being clear and specific, providing context, focusing on details, using iterative improvements, and including examples or templates. These practices lead to more accurate, relevant, and useful responses, demonstrating the importance of prompt engineering in maximizing the effectiveness of AI-driven solutions. Saving Changes...
I use structured techniques to refine and optimize prompts, ensuring clarity and alignment with the desired outcome. Some of the techniques I rely on are RFT (Role, Focus, Type), RAISE (Role, Audience, Information, Style, Examples), RIPEC (Role, Instructions, Purpose, End result, Client), and FOCUS+ (Focus, Objective, Context, User, Style, Output). Each serves different scenarios, from quick tasks to complex technical requirements.
For example, I once used RIPEC to refine a prompt in an energy project. The initial prompt was:
"Explain how to use AI in energy forecasting."
The result was too generic and lacked actionable insights. Using RIPEC, I reformulated it:
"Assume the role of an energy data scientist (Role). Provide step-by-step instructions (Instructions) for applying AI to hydrological data to forecast energy generation (Purpose). The end goal is to create a detailed report (End result) for energy company executives (Client)."
This change drastically improved the output, making it detailed, context-aware, and directly applicable to the project.
These techniques help me craft precise and actionable prompts that maximize the effectiveness of Generative AI
It provides more clarity and understanding of the query that we are looking for, otherwise only general responses will be received. Project Managers are too busy to execute their task efficiently and redefining the prompt will surely help them to extract the only required information with precise objectives. Well-refined prompts can lead to highly actionable insights that directly influence project planning, risk management, and decision-making processes. Saving Changes...
Ruba Abu SubaihTraining and Development Specialist, Project ManagerHouston, United States
I have noticed that when using AI, being highly specific, providing a clear persona, evaluating the output critically, and engaging in an ongoing conversation with the AI are crucial for achieving the best results. Saving Changes...
James TannerVice President| TurningPointSilver Spring, Md, United States
I first start with the KISS principle towards prompting and then I refine along the way. I also like using analyze and reference prompts to make sure I'm receiving appropriate responses. Saving Changes...
Thanks so much for these additional AI Prompt formulas.
I have RTF [role task format] and CREATE [character request examples adjust types of output evaluate] from the PMI online course. Your formulas are new to me:
_ BAB: Before After Bridge
_ CARE: Context Action Result Example
_ RISE: Role Input Steps Expectations
_ TAG: Task Action Goal
Except for BAB, yours seem to be pretty intuitive. Yet I think the RTF and CREATE forms encapsulate all the above. Thanks again!
Vanessa, I appreciate your succinct recap of the formulas. There is a lot of great info in PMI's AI-prompt Engineering course! Saving Changes...
There are framewoks to create prompt. This is part of the Prompt Desing discipline. Those that gave me and the initiatives where I was included are:R-T-F (Role-Task-Format), T-A-G (Task, action, goal), B-A-B (Before, after, bridge), C-A-R-E (context, action, result, example), R-I-S-E (role, input, steps, expectations).
Muchas gracias por este aporte, muy útil Saving Changes...
PMI Team Member
