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?
Thank you for the opportunity to participate. From my experience leading AI-powered digital products like BESTDoc, I believe there are several best practices that help ensure AI-generated results are accurate, relevant, and aligned with the original business goals:
Clearly define the objective. Before using any AI system, it's essential to have a precise understanding of the problem you're trying to solve. A well-structured question or workflow is the foundation of meaningful results.
Use high-quality, relevant data. The accuracy of AI depends heavily on the input data. Curated, representative, and up-to-date datasets are key to minimizing bias and improving outcome reliability.
Keep humans in the loop. Especially in sensitive or high-stakes contexts, AI outputs should be reviewed and validated by human experts before acting on them. Human oversight adds a critical layer of judgment and responsibility.
Iterate and monitor continuously. AI systems are not “set and forget.” Ongoing evaluation, performance tracking, and user feedback are necessary to maintain relevance and trust over time.
Contextualize results. Even if technically correct, AI outputs must make sense in the organizational, legal, or cultural context. Collaboration with domain experts ensures the output is both useful and applicable.
Promote transparency and explainability. Whenever possible, use interpretable models or explainability tools to make decisions traceable. This is especially important in public sector and regulated environments.
In short, AI can be a powerful enabler—but only when used with purpose, discipline, and responsible oversight. Our goal is not to replace human intelligence, but to amplify it.
Saving Changes...
Rajesh HaridasSolutions Manager, MBA, PMI-ACP, PMP, B.E., CSM, CSPONaperville, Il, United States
I found this interesting hack. Apparently, when using a persona, role, context, instructions, samples, constraints, and formats if you use markdown the LLM is able to return a better response.
for example, the following prompt.
Please research and identify the **Top 5 best vegan protein powders** available for purchase in San Jose.
Your evaluation must be based on a comprehensive analysis of the following criteria, and you must present your findings as a ranked list from 1 to 5.
**Evaluation Criteria:**
1. **No 'Protein Spiking':** The ingredients list must be clean. Avoid products with 'AMINO MATRIX' or similar proprietary blends designed to inflate protein content.
2. **Transparent Amino Acid Profile:** Preference should be given to brands that disclose a full amino acid profile, with high EAA and Leucine content.
3. **Sweetener & Sugar Content:** Scrutinize the ingredient list for all sugars and artificial sweeteners. For each product, you must **list all identified sweeteners** (e.g., sucralose, stevia, erythritol, aspartame, sugar).
4. **Taste Evaluation from Reviews:** You must search for and analyze customer reviews on US e-commerce sites (like Body & Fit, bol.com, etc.).
Summarize the general consensus on taste. Specifically look for strong positive reviews and strong negative reviews using keywords like 'delicious', 'great taste', 'bad', 'awful', 'impossible to swallow', or 'tastes like cardboard'.
5. **Availability in San Jose:** The products must be easily accessible to US consumers.
**Required Output Format:**
For each of the Top 5 products, please provide:
- **Rank (1-5)**
- **Brand Name & Product Name**
- **Justification:** A summary of why it's a top product based on protein quality (Criteria 1 & 2).
- **Listed Sweeteners:** The list of sugar/sweetener ingredients you found.
- **Taste Review Summary:** The summary of your findings from customer reviews. Saving Changes...
To ensure AI outputs are accurate and aligned with your goals, start by clearly defining what success looks like. Use layered prompting, iterating with context to refine responses. Always validate important content against trusted sources, and test for edge cases when applying AI to repeatable tasks. Framing prompts with roles or scenarios helps guide tone and relevance. Most importantly, maintain human oversight: AI can support thinking, but judgment, ethics, and leadership remain human responsibilities. The best results come from treating AI as a collaborative partner, not a shortcut. Saving Changes...
Carlos JazbinsekProject Management| Independent Professional ServicesRio de Janeiro, Brazil
As a standard procedure when dealing with AI, we should always check the outputs against the reality of the real world situation. As expected, a first output may be incomplete or imprecise which requires a good deal of prompt refinement until we can consider the output as close as we can accept. Saving Changes...
Godswill EgegwuAgile Coach, Scrum Master®, Author & TechpreneurAix-en-Provence, France
It’s my culture to always remember the basic concept of computing (Input, Process, Output), or simply GIGO (Garbage in, Garbage out - the quality of output is determined by the quality of the input). Hence, I always ensure I write prompts that are extremely clear, by stressing areas of importance while loading the prompt with as much relevant detail as possible related to my query.
I must confess, sometimes it takes a lot of time putting the prompts together because of how detailed I want them to be, but I'm sure the output will always be worth the time.
That's not all, I strongly believe the prompter has a post-responsibility in validating the outputs of the prompt, no matter how much detail, resources, or trust is involved.
That's a very good question. In my response, I am assuming that the question refers to an LLM-based chatbot.
From my experience, the best results are achieved the more context I provide to the LLM. This means providing as much information as possible that describes both the project itself and the project context.
A second very important step is the quality of the request, also known as the prompt for the LLM. This is similar to human communication, where the quality of the question determines the quality of the answer. Therefore, a good prompt strategy is required, for example:
1. Data and context about the project
2. The goal of the request
3. The task that the LLM should fulfill
4. The format in which the output should be delivered.
In subsequent requests, it is possible to build on the context and results of the previous request. It is important that this process takes place within a chat, as otherwise the context is lost.
i think your point of view is correct as strategy of thinking ,quality of information , understand the LLMs ,also continuous improve the questions to be advance will reflect on the output of the chat bot Saving Changes...
PMI Team Member
