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?
Akinwale AkinolaHead, Project Management| JNC International LtdSurulere, Lagos, Nigeria
Prior to being exposed to the structured approach to iterative refinement of prompts, I normally read outputs and refine them with further instructions which further fine-tune the responses towards a desired output. It works all the time and I used ChatGPT 4o. Saving Changes...
Vladimir QuinteroProfessor| Simon Bolivar UniversityBarranquilla, Colombia
Overall perceived quality of AI responses has to do with their usability. In turn, it depends on the use and specificity of three components:
1 Role: who are you asking to? Do not expect a precise/adequate response if you do note define this component.
2 Audience: who will use the response?
3 Tone and output format: what will be done with the response? Saving Changes...
To refine a prompt, it is important to consider the CREATE methodology: Character, Request, Example, Adjustments, Type of Output, and Evaluation. For large prompts, it is better to break them down into subtasks to ensure accurate AI comprehension and to provide clear feedback, reducing the likelihood of deviations in its responses.
Initial Prompt: "Describe a migration strategy for SAP S/4HANA."
Output: A general response with high-level steps.
Refined Prompt: "Create a detailed migration strategy for transitioning from SAP ECC 6.0 to SAP S/4HANA, focusing on data migration, system testing, and post-migration support, for a global media enterprise with 10,000 users."
Improvement: The output became more actionable, tailored to the specific enterprise size, industry, and post-migration needs.
The precision of the prompt is critical to obtaining actionable and intuitive results. A well-crafted prompt sets the stage for more relevant, targeted outputs, which is essential for any complex projects. Saving Changes...
I have used Generative AI for various complex tasks, such as developing detailed project plans and generating research summaries. Initially, the AI-generated responses were too broad and lacked the technical precision needed for my team, especially when dealing with highly specialized R&D projects. I have drastically improved the relevance and quality of the AI's responses by refining a prompt to include more context and specific details. For example, instead of asking the AI to create a project timeline, I refined the prompt to create a detailed product development timeline for a CLEANTECH research project, including key phases, milestones, resource allocation, and regulatory considerations. This prompt refinement allowed the AI to deliver a more detailed, tailored, and actionable plan. Adjusting the specificity and providing clear parameters leads to more accurate and valuable results, ensuring the output meets the project's exact needs. Saving Changes...
Refining a prompt is beneficial for the next prompt. A prompt refinement can add improvement and knowledge. Therefore, it is essential to document the prompt used and provide any metrics for its evaluation. Finally, it is advised to use the template for evaluation and documentation provided in the video. Saving Changes...
Anonymous
Refining a prompt in GenAI can significantly improve output quality by providing clearer context, reducing ambiguity, and narrowing the focus. A well-crafted prompt helps the AI understand the specific needs and constraints of the task, leading to more accurate, relevant, and targeted responses. Small adjustments in wording or structure can make a big difference in ensuring the AI’s output aligns more closely with the desired goals.
PMI Team Member
