Project Management

Please login or join to subscribe to this thread

In your experience with GenAI, how has refining a prompt drastically changed the output quality?

linkedin twitter facebook   Artificial Intelligence  
avatar
Sarah Philbrick
PMI Team Member
Director, Learning Design & Development| PMI Asheville, 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?
Sort By:
< 1 ... 75 76 77 78 79 80 81 82 83 84 85 ... 147 >
avatar
Tharindu Wijesinghe Sri Lanka
In my view when we ask something from Ai usually without using any method or a pattern it may not give us the actual output we need . But with methods like CREATE and Pattern link React we can have more interaction with the AI and can be asked more specifically what we required and AI also can learn and have a better understanding of our requirement when we use Iterative process as well and we can simplify our requirement by creating sub task as well so the output will be much quality meaningful .
avatar
Carol Robinson New Britain, CT, United States
Often, when I ask AI to summarize a document without specifying a persona or providing an example, the result tends to be very generic and not particularly useful. However, by refining the prompt to include more detailed information, the response becomes much more well-formulated and relevant.
avatar
Arnoldo Fernandez San Jose, Costa Rica
Providing a picture with a description of the expected input and output
avatar
Amit Lodha Senior Manager| Deloitte & Touche Middle East Abu Dhabi, Abu Dhabi, United Arab Emirates
Defining and refining has always helped in all situations and walks of life. While refining the prompts through iterative approach helped me get better results - more to the point and closer to the desired results, it also helped me become a better manager. I realized that this is identical to managerial instruction, the more precise and structure our instruction, better are the results - team delivers better results, learns faster, optimizes TAT and this also helps develop mutual respect in the team - happiness everywhere - Prompt Engineering helped me get better in life.
avatar
Chuan Chong Tan Singapore, Singapore, Singapore

core principle of effective AI interaction! Here are concrete examples demonstrating how specific prompt refinements drastically improved output quality, along with the key changes made:


Example 1: Image Generation (e.g., Midjourney, DALL-E)

Initial Prompt: "A modern office lobby"



Result: Generic, bland, low-detail image. Could be any corporate space.



Improved Prompt: "Award-winning architectural photography of a sun-drenched, minimalist Scandinavian-inspired office lobby in a tech company headquarters. Features natural wood accents, lush vertical gardens, ergonomic furniture groupings, a double-height ceiling with skylights, and a subtle abstract sculpture as a focal point. Soft, natural lighting, 35mm lens, depth of field, ultra-realistic, 8K."



Key Changes & Why They Worked:



Added Style & Quality Cues: "Award-winning architectural photography," "Scandinavian-inspired," "ultra-realistic, 8K" set the desired aesthetic and fidelity bar.



Increased Specificity: "Sun-drenched," "minimalist," "natural wood accents," "lush vertical gardens," "ergonomic furniture," "double-height ceiling with skylights," "abstract sculpture" provided concrete visual elements.



Defined Context: "Tech company headquarters" gave cultural/functional context.



Specified Composition: "35mm lens, depth of field" guided camera perspective.



Enhanced Ambiance: "Soft, natural lighting" set the mood.



Result: A highly detailed, visually striking, and architecturally coherent image matching a specific vision.
avatar
Haroon Ur Rasheed Director New Products and Delivery| MicroMerger Islamabad, Pakistan
Building a common understanding of the problem is the most important factor. Then, the results/output would be more concise and satisfying.
avatar
Panjari Hiteshkumar Mangalbhai Project Manager| GHCL Kutchchh, India
In my experience, refining a prompt can significantly transform the quality, clarity and relevance of the output generated by GenAI. A vague or generic prompt often results in superficial or broad responses that may not align with the specific project context. However, when I reframe the prompt to include precise objectives, constraints and desired formats such as timelines, industry specifics or stakeholder expectations, the output becomes far more insightful and actionable.
avatar
Belcon Francisco Sugar Land, Texas, 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 so much for sharing these frameworks!
avatar
Aruna Reddy Shine soft LLC Parker, CO, United States
concise and clear and give prompt with expected results and refine through practice.
avatar
Gonzalo Justiniani Senior Project Manager| PanCO Construction Consultants, Inc Panama, Panama, Panama
When I started using it, I had no previous experience with LLMs but quickly realized that in order to get the answers I needed, I had to let GenerativeAI from which poit of view I wanted it to analyze the situation. It is not the same to evaluate a problem from a contractor's point of view than to do it from an Owner's rep. point of view.
< 1 ... 75 76 77 78 79 80 81 82 83 84 85 ... 147 >

Please login or join to reply

Content ID:
ADVERTISEMENTS

You may have to fight a battle more than once to win it.

- Margaret Thatcher

ADVERTISEMENT

Sponsors







Vendor Events

See all Vendor Events

Newsletters

Jun 3: The Career Problem Facing New PMs
May 20: Your PM Career as a Project Life Cycle
May 13: The Big Lie After Layoffs

See all newsletters