Director, Learning Design & Development| PMIAsheville, NC, United States
Validating and checking outputs is critical when working with AI systems like Generative AI. Such validation approaches may include establishing clear criteria, implementing strong testing protocols, and continuous refinement.
In your experience with AI, what are some best practices for ensuring the results you receive are accurate, relevant, and aligned with your original goals?
It is important to keep record of the used prompts and the refinements. Adopt a prompting method like CREATE and keep checking the output, respectively provide feedback to AI where it failed to fulfill the expectations. Saving Changes...
for me in the experimental phase i will do it my self and then compare with the results that will be generated from AI till i get the confidence and accurate prompts writing and get the correct and accurate results Saving Changes...
Pranav LakhkarFounder| ProMiles Management ConsultingPune, MH, India
With a deep understanding of the method how AI responds to our queries or prompts, I believe below points are important in getting accurate and relevant results from AI tools.
1. start with clear, specific prompts and provide context.
2. Treat AI as a collaborator—iterate with feedback and break down complex tasks.
3. Always validate critical information and stay focused on your original goals.
4. Use structured formats (like bullet points or tables) for clarity, and avoid sharing sensitive data.
5. Understand AI’s limitations and use it as a drafting or ideation assistant, not a final authority.
6. Saving effective prompts for future use can also boost productivity.
While using AI, to get more accurate the relevant data request with specific context, with examples and if required provide the type of information and with details in structured format. Saving Changes...
Using a specific context and the format is key, but we always need to double check the results provided Saving Changes...
Janice TimchuckChief of Staff| Wayne State University School of MedicineSaint Clair Shores, Mi, United States
Jul 10, 2024 2:35 PM
Replying to Giovanni Alonso Alvarado Morales
...
To ensure the AI system results I receive are accurate, relevant, and aligned with my original goals, I will introduce the following best practices from the course material:
1. Iterative Refinement: I will continuously adjust prompts based on the AI's responses, providing additional context and necessary clarifications.
2. Clear and Specific Instructions: I will offer detailed and specific instructions in my prompts to avoid ambiguities.
3. Structured Formulas: I will use formulas like RTF (Role, Task, Format) or CREATE (Character, Request, Examples, Adjustments, Types of output, Evaluation) to effectively organize prompt components.
4. Validation Checks: I will implement validation by requesting citations and verifying information against known facts.
5. Providing Context and Eliminating Irrelevant Information: I will ensure sufficient context without overloading the model with unnecessary details.
6. Confidentiality and Ethical Use: I will handle sensitive data carefully and adhere to ethical guidelines.
7. Regular Feedback and Continuous Improvement: I will establish mechanisms for regular feedback and continuously improve AI performance.
This was very helpful! Saving Changes...
Janice TimchuckChief of Staff| Wayne State University School of MedicineSaint Clair Shores, Mi, United States
Jul 04, 2025 3:34 AM
Replying to Pranav Lakhkar
...
With a deep understanding of the method how AI responds to our queries or prompts, I believe below points are important in getting accurate and relevant results from AI tools.
1. start with clear, specific prompts and provide context.
2. Treat AI as a collaborator—iterate with feedback and break down complex tasks.
3. Always validate critical information and stay focused on your original goals.
4. Use structured formats (like bullet points or tables) for clarity, and avoid sharing sensitive data.
5. Understand AI’s limitations and use it as a drafting or ideation assistant, not a final authority.
6. Saving effective prompts for future use can also boost productivity.
From my experience, I believe it comes down to a combination of disciplined input, intelligent interaction and rigorous validation. Below are some best practices that I advocate:
1. Before interacting with an AI system, clearly articulate what you want to achieve, why you need the information and how it will be used within your project. This will help you to craft precise prompts and evaluate the output.
Provide ample, relevant context. AI models perform best when given sufficient background information. Don't just ask a question: provide the necessary details, constraints and stakeholders, as well as the specific project phase or problem you're addressing. The more context you provide, the more relevant the output will be.
'Quality in, quality out' (QIQO): This familiar data management principle also applies to AI. If your source material is flawed, the AI's output will likely be too.
2. Avoid vague prompts. Instead of 'Tell me about the risks', try 'Generate a list of potential technical risks for a cloud migration project to AWS, considering data security and GDPR compliance, in table format with columns for risk, likelihood, and impact.' If the initial response is not perfect, refine your prompt. Think of it as an agile process: prompt, review, refine, re-prompt.
Set constraints and formats: State the desired output format explicitly (e.g. bullet points, table or executive summary), as well as the desired length, tone and any specific elements that should be included or excluded.
Test assumptions and explore nuances. Don't just accept the first answer. Ask follow-up questions to probe the AI's understanding, challenge its assumptions and explore different facets of the problem. For example: 'What are the counter-arguments to that approach?' or 'Can you provide a different perspective?'
3. Human-in-the-loop validation: Never blindly accept AI output. Cross-reference information with reliable, independent sources. Use your own domain expertise, critical thinking and project knowledge to verify the accuracy and relevance of the information.
Align to goals and constraints: Review the AI's output against your initial objectives and all known project constraints, such as budget, timeline, resources and regulatory requirements. Ask yourself: does it truly help move the project forward, or does it introduce new complexities or misalignments?
Contextualise and refine for your audience. AI output is often generic. As the project manager, you must tailor, refine and contextualise it for your specific project team, stakeholders and organisational culture.
Document and learn: Just as we do with lessons learned from projects, it is beneficial to document what worked well and what didn't when using AI for specific tasks. Saving Changes...
The more precise information (quality of the data) is provided and the request for what is wanted (the output) is well defined, the clearer the answer we will get will be and in line with what we expect. That is why the structure that is put together when making the query is of utmost importance, where it is tried to cover the greatest amount of information to obtain the desired output. Saving Changes...
PMI Team Member
