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?
Ganesh KanagasabapathyProject Manager| Asasra Teknik Sdn Bhd24, Jln 10, Taman Batu, 68100 Batu Caves, Selangor, Malaysia
I use prompt chaining, which involves breaking large tasks to sub-tasks to complete one or 2 task at a time and creating subsequent prompts to solve multiple problems at once. Reason is an overloaded prompt the AI is only able to complete 2 of the 4 tasks or only able to provide a very general answer so a chain prompt with CREATE formula required. We sustain with the role but give a specific request and follow the same method to the following tasks and sub task. Each tasks is constructed as a prompt. It will a better, easy to read and comprehensive. It allows PMs to give a better and more targeted feedback to AI.
PMI prompt documentation template is available. To maximize effort and crafting an effective prompts, documentation becomes a fundamentals component. This increase efficiency in the AI model. Saving Changes...
The output changed from generic content to a structured, business-focused presentation tailored to the audience, with clear messaging and usable slides.
If I give a vague prompt like “summarize this document”, the output is often too general and may miss important points.
When I improve the prompt by specifying:
what kind of summary is needed (technical, commercial, risks, etc.)
the level of detail (short summary or detailed points)
the format (bullet points, table, etc.)
focus areas (e.g., only key requirements or deviations)
the result becomes much more useful and directly applicable to my work.
Saving Changes...
RAUL PONCE DE LEONCEO| SERVICIOS INTEGRALES DE VENTA EIRLMagdalena, LIM, Peru
In my experience, refining a prompt is the difference between receiving a 'generic template' and a 'tailored solution.' I've seen how a vague instruction produces 'hallucinations' or shallow summaries, whereas a refined, high-precision prompt incorporating context, persona, and specific constraints drastically shifts the output toward actionable intelligence. However, this 'fine-tuning' must be balanced with strict data responsibility. In the quest for better quality, we must never feed confidential project data into public models. Mastering the art of the prompt also means mastering the art of anonymization, learning how to provide enough context for the AI to be effective without compromising our organization's proprietary information. Ultimately, a refined prompt isn't just about better text; it's about mitigating risk and ensuring the output aligns with professional standards. Saving Changes...
Rakesh JhaManagement| Mohur Fintech Private LimitedBengaluru, Ka, India
It is essential to provide clear context when working with any task, as it helps guide the response toward your intended goal. Without sufficient background or direction, the outcome may not fully align with what you are looking for. Equally important is the willingness to refine the process—if the initial result does not meet your expectations, you should not hesitate to ask for adjustments or explore alternative outputs. Iteration plays a key role in improving the quality and relevance of the final result. Additionally, maintaining consistency and perseverance throughout the process is crucial. Achieving the desired outcome often requires multiple attempts, careful evaluation, and continuous fine-tuning. By staying patient and committed, and by clearly communicating your needs, you significantly increase the chances of obtaining a result that truly matches your purpose and expectations.
Saving Changes...
Wayne HoughtonStrategic Programme Leader| Retail & Financial Services IndustryCape Town, Western Cape, South Africa
I have seldom had a prompt achieve success on the first round. I think about prompt engineering a bit like talking to another human. Even with humans we have to clarify and have dialogue back and forth to ensure we have correctly understood each other. Even when requesting a team member to complete a task requires some refinement and iteration of the task, reviewing the output and giving feedback. With AI, the turnaround is just alot faster. Saving Changes...
It is important to remember this: generative AI is just "predictive text with storoids". Obviously not only text will be the result. BUT the important thing is the answer will just to complete your question (prompt) with the things that have more probability to complete it. You can manage it using some of the parameters like temperature. So, it is very important when creating the prompt to put clear the role, the place where the role works/live/etc, the task the role has to accomplish and the format of the answer. This is an example of R-T-F. You have to eliminate as ambiguity as possible. If not, then hallucinations will happened.
Thank you for the insight. I experienced generative AI works best with R-T-F when we create the persona and keep refining the answers mentioning the same role. AI also suggest refinements of the result based on saved history and how you represented chain of thoughts in past. Analyzing the results is something I still need to explore but considering parameters like the place we live, weather and cultural details can help bringing the detailed answer. Regular feedback can avoid the frequency of hallucinations.
In my experience, refining a prompt has consistently been the difference between a generic response and one that is immediately actionable.
Early on, broad prompts produced broad answers — technically correct but rarely useful. As I became more deliberate about including context, constraints, and desired format, the outputs became targeted and required far less rework.
In a project management context this matters because acting on misaligned information carries real consequences. A well-crafted prompt that specifies the audience, the decision being made, and the format needed produces a response I can use directly — saving time and reducing stakeholder miscommunication.
The core lesson is simple: the precision you put into a prompt is directly proportional to the value you get out of it.
In my experience, refining a prompt has consistently been the difference between a generic response and one that is immediately actionable.
Early on, broad prompts produced broad answers — technically correct but rarely useful. As I became more deliberate about including context, constraints, and desired format, the outputs became targeted and required far less rework.
In a project management context this matters because acting on misaligned information carries real consequences. A well-crafted prompt that specifies the audience, the decision being made, and the format needed produces a response I can use directly — saving time and reducing stakeholder miscommunication.
The core lesson is simple: the precision you put into a prompt is directly proportional to the value you get out of it.
It is important to remember this: generative AI is just "predictive text with storoids". Obviously not only text will be the result. BUT the important thing is the answer will just to complete your question (prompt) with the things that have more probability to complete it. You can manage it using some of the parameters like temperature. So, it is very important when creating the prompt to put clear the role, the place where the role works/live/etc, the task the role has to accomplish and the format of the answer. This is an example of R-T-F. You have to eliminate as ambiguity as possible. If not, then hallucinations will happened.
PMI Team Member
Community Champion
