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?
Chanel BosticSenior Project Manager| AARPPembroke Pines, Fl, United States
great
Saving Changes...
Sam WoottonCEO| Quark ConsultingFulshear, Tx, United States
Not surprisingly, a "little" knowledge (e.g. prompt structure and models) has exponentially improved responses. Looking forward to working with chaining and Q&A more. Saving Changes...
In my experience using GenAI, refining a prompt drastically transformed output quality by shifting the model from producing generic, surface-level responses to delivering targeted, decision-ready insights. Early prompts that were broad or ambiguous often resulted in verbose but unfocused outputs. By contrast, specifying the role, constraints, frameworks, and success criteria consistently produced outputs that were more accurate, relevant, and immediately usable. This mirrors good project requirements definition, clearer inputs reduce rework and ambiguity downstream.
From a project management perspective, prompt refinement functions like scope clarification and stakeholder alignment. Iterative improvements, adding context, assumptions, and deliverable format, reduce variance and risk in AI-generated results, much like progressive elaboration in project planning. High-quality prompts also improve traceability and ethical use by making expectations explicit, enabling the PM to validate outputs against objectives rather than accepting them at face value.
In my experience using GenAI, refining a prompt drastically transformed output quality by shifting the model from producing generic, surface-level responses to delivering targeted, decision-ready insights. Early prompts that were broad or ambiguous often resulted in verbose but unfocused outputs. By contrast, specifying the role, constraints, frameworks, and success criteria consistently produced outputs that were more accurate, relevant, and immediately usable. This mirrors good project requirements definition, clearer inputs reduce rework and ambiguity downstream.
From a project management perspective, prompt refinement functions like scope clarification and stakeholder alignment. Iterative improvements, adding context, assumptions, and deliverable format, reduce variance and risk in AI-generated results, much like progressive elaboration in project planning. High-quality prompts also improve traceability and ethical use by making expectations explicit, enabling the PM to validate outputs against objectives rather than accepting them at face value.
Saving Changes...
Afhaz AhmedSr Service Delivery and Program Management| Bahwan CybertekMuscat, Oman
Another Great question @Sarah. In my experience, refining prompts works almost exactly like refining user stories or technical requirements in a development project—the clearer and more structured they become, the better the final output.
Here are a few examples of how improving a prompt drastically changed the quality of results:
1. From vague requirement to structured technical specificationInitial prompt: “Summarize the project status.”
This produced a generic summary with missing details—similar to giving a developer a vague requirement and expecting a perfect feature.
Improved prompt: “Summarize the project status for the UAT phase, including the number of test cases executed, defects raised, blockers, upcoming milestones, and any vendor dependencies.”
Impact: The output shifted from broad and high‑level to a structured, accurate, and manager‑ready summary—just like refining acceptance criteria in a user story.
2. Converting a casual request into a development‑style task descriptionInitial prompt: “Create a risk log.”
This gave a simple table with generic risks.
Improved prompt: “Create a detailed risk log for an integration project with OPAL. Include technical, timeline, and dependency risks, probability, impact rating, mitigation actions, and responsible owners.”
Impact: The new output resembled a proper project artifact—much like when developers get clear technical context instead of assumptions.
3. Adding constraints to reduce “AI hallucinations”Initial prompt: “Explain the API integration approach.”
This produced a theoretical explanation with concepts not used in our environment.
Improved prompt: “Explain the API integration approach using REST-based architecture, JSON payloads, token-based authentication, and structured error handling aligned with our standard middleware process.”
Impact: By adding constraints, the output became aligned with our actual architecture—similar to defining coding standards so developers don’t go off‑pattern.
4. Defining the audience—just like tailoring communication for executives vs. developersInitial prompt: “Write a progress update.”
It produced a technical-heavy update.
Improved prompt: “Write a progress update for senior leadership, focusing on milestones, risks, mitigations, and business impact. Keep technical details minimal.”
Impact: The tone and structure became executive‑friendly. In project management terms, this is like refining communication plans based on stakeholder expectations.
Overall takeawayJust like in the programming world, better requirements produce better outputs.
Every time I refined the prompt with:
Increasing specifity and more context leads to more accurate and refined results/output
Saving Changes...
Haruna BulusMr.| Abuja Electricity Distribution CompanyAbuja-Fct, Fct, Nigeria
Some responses may seem vague initially, but once I refine my prompt, they suddenly reveal a treasure chest of information concerning the subject in question.
A well‑crafted prompt removes ambiguity, and iterative refinement — including chained prompting — ensures the AI produces accurate, relevant, and coherent outputs that align with goals.
PMI Team Member
