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?
Being concise and specific helps the AI to give some valuable answers. It also learns with time as you ask further questions.
I couldn't agree more. You need to be specific and constantly reiterate to produce better results Saving Changes...
Bhuvaneswari NatarajanPDI Data & BI Integration Architect| Shell India Markets Private LimitedChennai, India
Jun 21, 2024 7:28 AM
Replying to Sergio Luis Conte
...
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).
Great list of prompt design frameworks! I’ve found R-T-F and R-I-S-E especially helpful for structuring prompts in data and BI contexts. Curious—do you find some frameworks work better for creative vs analytical tasks? Saving Changes...
When I was preparing for my PMP exam, I relied on ChatGPT, Copilot, and PMI Infinity to validate my answers. Interestingly, the responses from the three tools often differed. I noticed that PMI Infinity stood out—it wasn’t generating random answers. While it likely leverages GPT, I believe it also uses RAG with PMI’s knowledge base, as it consistently referenced PMI materials. Still, even with Infinity, I achieved around 80% accuracy only after actively engaging with it by asking counter questions, probing with “why” and “why not”, to refine its responses. That’s when it struck me: prompt engineering isn’t optional, it’s essential. If you are using the LLMs for exam prep, please provide additional details (using RTF/CREATE Approaches) instead of copy-pasting the questions from the practice tests. For Example: Just by adding "Answer following PMI PMBOK Guidelines" to your prompt , you would see a drastic difference in the output. Saving Changes...
Anonymous
Until now we have not structured our prompts using any frameworks and ultimately led to average results and lack of AI adoption across the team. Now using the frameworks above and RTF and CREATE I expect better success and more team adoption Saving Changes...
Refining a prompt improves output by clarifying intent, adding context, and guiding tone or format. Specific prompts reduce ambiguity, leading to more accurate, relevant responses. Even small adjustments—like rephrasing or adding examples—can drastically enhance quality, turning vague answers into tailored, insightful results that better match user goals. Saving Changes...
Timothy KenneyProgram Manager| USAFBlack Hawk, Sd, United States
This course on prompt engineering gave me an idea for a promt formulation tool. The concept is to use the CREATE method (Context, Role, Examples, Ask, Tone, Evaluate) as a structured worksheet that helps us build better prompts outside of the AI tool. Once filled out, the worksheet generates a ready-to-use prompt that can be copied directly into ChatGPT or any other AI platform.
I think this could really help our project managers start using AI more effectively in their day-to-day tasks—whether it’s analyzing resource allocations, drafting communications, or building reports. It gives a repeatable process instead of trial and error, and ensures our AI interactions are purposeful and productive.
I’ve already started prototyping a simple web app version. Happy to share the link and get your feedback if you’re interested!
In my experience, Chain Prompts helps to gradually shape the outcome. It's similar to conversation with a person. In this video I learnt that if we provide complex details at-one-go, even machines also gets to confusion/ hallucinations. So keeping things simple makes life easier. Saving Changes...
Outcome of rewritten prompt has more specific details compared to First prompt. Always assume that you need to provide relevant details for better outcome.
Initial Prompt - Create a Business strategy for customer retention.
Rewritten Prompt - Act as a globally recognized IT business leader with over four decades of experience in the IT services industry, specializing in enterprise solutions and client relationship management. Your task is to develop a comprehensive customer retention strategy for a Fortune 100 company serving a diverse, worldwide client base. The strategy should address current disruptive challenges including AI-driven digital transformation, economic instability, and rapidly evolving customer expectations. Provide a detailed and prioritized list of actionable customer retention initiatives, clearly ranked based on ease of implementation and potential business impact. Include strategic rationale for each initiative and ensure the recommendations are scalable across regions and adaptable to different industries.
In my experience with GenAI, refining a prompt has often been the difference between a vague, unusable response and a highly actionable output. Applying structured frameworks such as R-T-F (Role–Task–Format) or C-A-R-E (Context–Action–Result–Example) helps ensure clarity, precision, and alignment with project goals. For example, when I shifted from a general task description to a C-A-R-E prompt, the output quality improved significantly—moving from generic insights to well-structured recommendations directly applicable to stakeholder needs. Prompt design is, therefore, not just a skill but a discipline that enhances project value delivery. Saving Changes...
"Humanity has advanced, when it has advanced, not because it has been sober, responsible and cautious, but because it has been playful, rebellious and immature."
PMI Team Member
