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?
from my point of view i used very short request to avoid overlapping or missing some tasks and kept it iterative to refine the final request Saving Changes...
Anonymous
In my experience of using GenAI for managing corporate projects to managing non-profit community programs, I have learned that refining the prompts with specifics, examples and evaluation is very important. The output will be more aligned to your needs and goals. Saving Changes...
Linn Chee HomProject Manager| State of CaliforniaSacramento, Ca, United States
I have not used GenAI yet. Will provide feedback/comment when I do. Saving Changes...
Emily SoleilGrant & Special Projects Manager| Harbor CareNh, United States
I find that continually discussing things with the AI and asking it questions like "what am I missing" or "is there value to doing it this way" or "Why did you choose to change or adress that next" is helpful in understanding the way the AI is "thinking" about things, or maybe it's understanding the context it's operating under. I just learned about feeding thigns back to it, so I've started feeding back the things that I pull out and edit, and we'll see how it improves my GPT. Saving Changes...
Santosh GiriFunctional Manager| WSP Consultants India Private LimitedBangalore, India
As better understanding of the question provide easy of finding the solution, creating smart prompt provide GenAI to understand the question easily and respond with relevant content. Saving Changes...
Prompting is part of discipline. GenAI helped me a lot in my day to day Project Management Role. With this prompt program will that provide useful output. Saving Changes...
YU-LING HSIAOProject Manager| OmniHealth Group Inc.Taiwan
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).
I agree with these frameworks, they are really useful. If you write prompts with a clear structure, the AI will respond more accurately and less likely to go off topic. It really helps! Saving Changes...
Scott TheusSenior Project Manager and Agilist| BWX TechnologiesEuclid, Oh, United States
I'm coming late to the conversation and would like to share an observation. While I've admittedly not read all of the posts and replies in this thread I've noticed that the majority of the discussions focus on the format of the prompt like R-T-F (Role, Task, Format) and others, but not the sources of the data. If the prompt does not include constraints on the data sources the accuracy and timeliness of the results may be impacted.
For example, Booma Pugazhenthi shared the prompt "Generate ten different questions about project management, similar in style and difficulty to the PMP exam, and provide detailed answers for each." In this context, the GenAI model will search all sources of data and may create questions and answers that are no longer relevant or even correct under the current edition of the standard.
To correct this there needs to be a constraint on the sources of the data:
"Using information from PMI's PMBOK 6th edition or later and publicly available sources published between January 1st, 2020 and today generate ten different questions about project management, similar in style and difficulty to the PMP exam, and provide detailed answers for each."
When I use GenAI to write practice questions for my students I also prompt for explanations for the answers to the questions and specifically exclude Reddit as a source. After some iterations this is the prompt I currently use:
"Use information from PMI's PMBOK 6th edition or later and publicly available sources published between January 1st, 2020 and today. Do not use Reddit as a source. Generate ten different multiple-choice questions about project management, similar in style and difficulty to the PMP exam. Provide detailed explanations for each possible response to the questions, including why the correct answer is correct and why the incorrect answers are wrong. Cite your sources."
(Side note...I also tell my students to exclude Reddit when they use GenAI to answer a practice question they are struggling with. Reddit and other discussion platforms are great tools to bring people together for collaborative learning, but by their nature they will have both right and wrong answers. The models do not separate theses out well; if a question is posted and 30 people agree that an answer is correct even though it is not and only one person provides the correct answer and explanation then the model is more likely to return incorrect information to the prompt.) Saving Changes...
to identify and analyse project risk, it is interessing to use prompt chaining. Begin with a prompt to identify the risks and then with an exemple you can ask to AI to analyse the identified risk and give them cotation to priorize them Saving Changes...
PMI Team Member
