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?
Providing a clear and specific goal, enough context, and persona. Critically reviewing the outputs and iterating with additional prompts to refine the results. AI should be treated as associate to support us and we need to validate outputs before taking the final outcome.
Saving Changes...
Anonymous
I think user review of the results is of the upmost priority. Certainly the ability to do iterative prompting is vastly important as well.
To ensure AI outputs are accurate, relevant, and aligned with original goals, it is essential to clearly define the objective, expected outcome, and constraints upfront. Providing sufficient context, stating assumptions, and breaking complex requests into structured steps significantly improves output quality. Asking the AI to explain its reasoning and highlight uncertainties helps validate results, while iterative refinement allows misalignment to be corrected early. Finally, critical outputs should always be reviewed and cross-checked by humans, as AI is a decision-support tool rather than a source of accountability.
To ensure AI outputs are accurate, relevant, and aligned with original goals, it is essential to clearly define the objective, expected outcome, and constraints upfront. Providing sufficient context, stating assumptions, and breaking complex requests into structured steps significantly improves output quality. Asking the AI to explain its reasoning and highlight uncertainties helps validate results, while iterative refinement allows misalignment to be corrected early. Finally, critical outputs should always be reviewed and cross-checked by humans, as AI is a decision-support tool rather than a source of accountability.
Saving Changes...
Doga ilterProject Manager| ÜlkerBa?C?Lar, ?Stanbul, Türkiye
I like to reate and agent and then build the final output step by step. I also ask it to providesources or "cite" where is it is pulling it's information from. Breaking up the prompt into smaller segments seems to work well.
Saving Changes...
Amit Jain BarjyatyaFunctional Manager| Harman Connected ServicesBangalore, Karnataka, India
Jun 08, 2024 6:40 AM
Replying to Oliver Chitsamatanga
...
A very good question and also difficult to answer as well. However you have to go to the basics and say as far as you are concerned, how well are you versed with the subject at hand ?. There are facts which the AI will generate and if you can verify these facts the more reliable the generated response will be. The fewer the facts then it means that the Generative AI response is far from meeting your original goals. Then it becomes very critical that you review the accuracy , relevancy and the alignment of the response to your original need. Unfortunately there are no clearly defined metrics that one can use a model to evaluate an AI generated response. So from my personal experience I basically restrict AI to an area where i have sound knowledge of , else it becomes almost impossible to verify details generated by an AI if you venture into unchartered territory. However with long usage and exposure your confidence also tend to increase as well. The best practice and protocol to follow would be to consult subject matter expects to validate the AI generated response before making critical decisions based on it to avoid any inherent associated risks which you might be not aware of.
Correct and precise input with available support documents will help to get better outcome.
With sensative data, you need to generalize and provide the available document to be compliant with company policy and generic outcome can be used to align with your requirements.
Saving Changes...
MAHANTESH MUGALILead Project Manager| CaterpillarBangalore, Ka, India
In my view, the key to getting meaningful results from AI systems lies in how we engage with them. I always start by giving clear context—explaining the user persona, why the outcome matters, and what kind of response is expected. I also ask the AI to cite its references or provide my own when possible, ensuring confidentiality and privacy are respected. Equally important is the mindset: I treat AI outputs as a starting point, not the final word. That means evaluating responses critically, iterating with follow‑up prompts, and applying structured prompting techniques like CREATE, RTF, chain of thought, and chain of feedback. This approach helps me investigate correctness, refine relevance, and align results with my original goals. Ultimately, AI becomes most valuable when we use it as a copilot—guiding, questioning, and shaping its outputs—rather than passively accepting them. Saving Changes...
If you want reliable results from AI, treat it like a junior analyst, not an oracle. Be clear about the outcome, audience, and constraints before you prompt. Vague inputs produce generic outputs. Define the decision the output should support. Provide context such as project stage, risks, and assumptions. Break complex tasks into steps instead of asking for everything at once. Request sources for factual claims and verify critical information independently. AI drafts. You validate. Finally, evaluate whether the output actually advances your project goals or just sounds good. Alignment with objectives is the real metric. Saving Changes...
Paolo SalaSenior Project Control ManagerCDMX, Mexico
I talk to the AI tool like I'm talking to somebody that might help. If I'm not specific, AI answer won't be either. If I provide details, AI can give me a better answer. Then I have to adapt the answer to waht I need, and double check. The same happens when we talk to people. Saving Changes...
PMI Team Member
