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?
Refining a prompt can drastically improve AI output by breaking a large task into smaller more manageable chunks. For example, if creating a document, I will work section by section, then review and refine each output before moving on. Leading to a more focused and useful final result. Saving Changes...
Abi GabrielProgram Manager| AptivBangalore, Karnataka, India
In my experience, I see that if we refine the prompt, it gives a better response.
For most of the prompts, I follow the RTF: Role Task Format formulae.
Sometimes we can see there is a hallucination, in such instances – I ask to cross-check the data point and share siting references too – then I’ve seen the LLM correct itself, and provide a better/accurate response.
I also ask LLM to share concisely – by prompting to share 3 top tips/ideas/issues/critical items – so that I can focus on the key aspects.
At other instances, I’ll ask to tabulate with some categorization, etc.
I also practice to break complex tasks into sub-tasks so that we can each time have LLM focus on specific area and it enables to drive better results
I see overall the output quality of the prompts increases as we refine – it is more of a continuous improvement. Saving Changes...
A refined prompt gives the model clearer constraints, richer context, and a defined target — which dramatically improves accuracy, relevance, and depth. Saving Changes...
Sarvesh Kumar UpadhyayEmbedded QA Engineer| Interel Building AutomationDubai, DU, United Arab Emirates
In my experience, the difference between a rough prompt and a well-refined one is honestly huge.
When the prompt is vague, the output feels like a general template—safe, but not very useful. It usually misses context, goes off in the wrong direction, or gives very surface-level answers.
But when you refine it and add a bit more clarity—like what exactly you want, the situation, the format, or the role—the output suddenly feels much more “on point.” It starts sounding like it was written for your exact use case instead of for everyone.
For example, instead of saying:
“Write a QA bug report”
If you say something like:
“Write a QA bug report for a login issue on Android, include steps to reproduce, expected vs actual result, and keep it in Jira format”
The difference is immediately noticeable. Now it’s structured, relevant, and something you could actually paste into a ticketing system with almost no edits.
What I’ve noticed is that the model doesn’t really “improve” on its own—the clarity you give it is what drives the quality. The more specific you are, the less it has to guess, and guessing is usually where things go off track. Saving Changes...
In my experience, prompt refinement significantly improves the value of GenAI outputs. Each refinement helps increase accuracy, reduce irrelevant information, and produce results that are better aligned with the intended objective Saving Changes...
Input data and the structure of the data is key to creating a strategic prompt. Also allow for the AI to digest smaller amounts of structured data to create a clearly refined result.
PMI Team Member
