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?
I started using LLMs with a simple framework of Who-What-Why, where Who stands for the persona that the model has to assume, What refers to the task to be done, Why describes the context for which the user is seeking the particular action. Saving Changes...
Manuel OspinaDirector PMO| Expert Projects S.A.S.Santiago De Cali, Valle Del Cauca, Colombia
First Instruction (General) - Mediocre initial result:
Instruction:
“Generate a project schedule for software development, including major phases.”
Result:
Design: 1 month
Development: 3 months
Testing: 1 month
Deployment: 1 week
Analysis of the result:
Although the phases are present, the timeline is extremely generic and does not take into account the specifics of the team, resources, or potential risks. There is insufficient detail on the individual tasks within each phase, nor are the dependencies between them, or important milestones included.
Second Instruction (More Specific):
Refined Instruction:
“Generate a detailed schedule for a software development project. The project includes the design, development, testing, and deployment phases, with a team of 6 developers, 2 designers, and 1 tester. The design should include user research, wireframes and visual design. Development should be subdivided into backend and frontend, and testing should include functional and integration testing. Consider that the project has a 6-month limit and frontend development cannot begin until the visual design is complete.”
Refined result:
Design:
User research: 2 weeks
Wireframes: 2 weeks (dependent on user research)
Visual design: 3 weeks (depends on wireframes)
Backend Development: 2 months (parallel to visual design)
Frontend development: 1.5 months (after visual design)
Testing:
Functional testing: 2 weeks (parallel to frontend development).
Integration testing: 1 week (after functional testing)
Deployment: 1 week
Refined result analysis:
The new timeline is much more detailed and specific. It not only includes the subdivision of tasks within each phase, but also takes into account the dependencies between tasks (such as the relationship between visual design and frontend development). Time and resource constraints are also considered, making it a much more useful schedule for project planning and tracking.
Changes made for improvement:
Increased specificity in tasks.
Team context and resources.
Dependencies between tasks.
Time constraints.
Specify the tests. Saving Changes...
Refining prompt adjusts output from generalized response version to more focused on the context and additional details shared. This helps in lot of ways, User gets more precise and accurate response from AI after refining and fine tuning the prompt and Gen AI learns from this refining process by feedbacks and adjusted chain of prompts. Saving Changes...
Helena GenerPM I| Setting Consultoria en T.I., S.L.Barcelona, Barcelona, Spain
Specify the final audience, to specify the format, the language to use, to be more specific changes 100% the quality of my answer Saving Changes...
As a project manager using Generative AI, I improved output quality by refining prompts to be more specific and context-driven. For example, when generating a project timeline, the initial prompt, “Create a project plan,” yielded a generic outline. I revised it to include key details: “Create a detailed project plan for a 6-month software development project, including milestones for design, development, testing, and deployment.” This refinement drastically improved relevance, as the AI produced a tailored, step-by-step timeline with specific tasks, deadlines, and resource allocation. The key improvement was adding clarity and context to guide the AI toward more precise outputs.
Best regards.
Saving Changes...
Isaac MartinezProgram Support Specialist| RaytheonWorcester, MA, United States
As I mastered the refining process, I was able to see the following benefits...
• increased precision
• improved relevance
• greater depth
• and better overall contextual outputs Saving Changes...
Anonymous
It has turned vague responses into highly detailed usable responses. Saving Changes...
By refining prompts, one can significantly improve AI-generated results, achieving greater relevance, accuracy, and overall quality.
Example is related to Technical Documentation
Original Prompt: "Explain how to set up a VPN."
Initial Output: Overly technical, confusing instructions.
Improved Prompt: "Create a step-by-step guide for non-technical users to set up a VPN on their Windows laptop, including screenshots and troubleshooting tips."
Changes:
- Targeted audience (non-technical users)
- Added visual aids (screenshots)
- Included troubleshooting
Output Improvement: Clearer, more accessible instructions.
So Key Takeaways are:
1. Specify target audience and context.
2. Clarify objectives and tone.
3. Provide relevant details and constraints.
4. Refine language and phrasing.
5. Iterate and adjust based on output. Saving Changes...
PMI Team Member
