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?
Ishwar SinghProject Management| Kyndryl Solution Pvt LtdGhazibad, Uttar Pradesh, India
I think it is important to give the context and also to refine, asking for a different output in case that the first one is not completely suitable to our purpose or to the outcome that we were looking for. I think that consistency and perseverance in looking for the result, is crucial as well. Saving Changes...
Asif KhanProduct, Program, Customer and Executive Mgmt| IBMAshburn, Va, United States
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).
By adopting disciplines approach to prompt engineering and experience with different LLMs and especially documenting the results for the Integrated Product Teams (IPTs) awareness and engagements of other functional stakeholders such as contracts, engineering is key. It is a learning process where everyone need to buy in the results with confidence. Saving Changes...
Anonymous
Refining a prompt transforms vague responses into sharp, targeted insights.
In my experience, a well-tuned prompt can cut task time in half and deliver exactly what’s needed—no guesswork, no do-overs. Saving Changes...
abiodun omotayoTechnical Planner| PrivatePlantation, FL, United States
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).
Thank you for providing other prompt framework/formular that suite different task to get better output from using AI Saving Changes...
Anonymous
Users viewing AI as a means of not thinking through what they are looking for are misguided and leads to poor prompt writing. All good prompts require clear instruction. When training junior PMs I guide them through a series of questions that lead them to discover the answer, the same approach is required with AI. Saving Changes...
Zamir BradfordProgram Officer| de Beaumont FoundationAcworth, GA, United States
I’ve had countless instances where refining a prompt completely transformed the quality of the output. Early on, I often rushed—offering minimal details, skipping context, failing to assign the AI a role or tone, and expecting excellence from vague direction. Predictably, what I received was unusable—proof of the old adage, “trash in, trash out.” The turning point came when I recognized that the issue wasn’t with the model but with how I was communicating my intent. Once I began treating prompts as structured briefs rather than casual queries, the results improved dramatically. I started specifying the AI’s role (for example, “act as an experienced editor” or “a senior project manager”), outlining the purpose of the task, and including examples, tone preferences, and contextual details. Sometimes I’d refine the prompt progressively—iterating based on what worked and what didn’t. These adjustments consistently elevated the quality, nuance, and accuracy of the output. The experience reinforced a powerful truth: the clarity, depth, and precision of the input are directly proportional to the sophistication of the response.
Saving Changes...
Salah ZureikatMEP Project Manager| Hiba Engineering EstablishmentAmman, JA, Jordan
In my experience, the shift from poor to excellent GenAI output comes down to three things: specificity, recognizing the predictive nature of the tool, and using structured frameworks.
Be Explicit and Contextual: As Kiron demonstrated, simple language isn't enough. We must increase the specificity and detail in our requests. This means providing ample context and asking for the style/level of detail required (e.g., 'similar to the PMP exam'). The AI cannot read our minds; it can only respond to what we tell it.
Recognize the Predictive Engine: As Sergio Luis Conte highlights, Generative AI is a 'predictive text with steroids.' It finds the most statistically probable next text. Therefore, our prompt must eliminate ambiguity by clearly defining the role, task, audience, and format of the desired answer.
Use Frameworks (e.g., R-T-F): To avoid missing key instructions, I recommend adopting a simple prompt framework, like the R-T-F (Role, Task, Format) model, or a similar structured approach. This consistency helps the AI learn and provides us with a reliable method for generating better, more valuable answers over time. Saving Changes...
Refining prompts involves thorough thought regarding the tasks specifications, constrains, hierarchy, as well as the role needed to address them. Improving our awareness of the issue at hand as well as the quality of the AI's output. Saving Changes...
PMI Team Member
