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?
Brian EvansResearch and Development Director| Accreditation Commission for Health CareNC - Raleigh, NC, United States
You have to define the goals first. If you don't clearly define the goals to AI, then the responses are going to be vague and unclear. The definition comes through using the CREATE formula. The more information you give the tool, the better it will respond.
Saving Changes...
Patrick HeaneyAgentic Governance Architect | Transforming PMO into AI-Driven ValueStreams| NoneChapel Hill, Nc, United States
What really magnified the AI value generation for me was moving off Web and Applications UIs for LLMs and into Integrated Development Environments (IDEs), Command Line Interface (CLI) UIs for LLMs, and a touch of python. This combination really helps you organize, mature, and reuse your prompts. See, https://github.com/AVS-Orchestration/. Saving Changes...
Anonymous
Making your prompt as specific as possible, while also sharing examples or references to tailor its response.
Yes I agree with most of the answers, have a clear prompt engineering, pattern, continue feedback the model, keep the context, use other LLM models or AI prompts, validate with SME and apply the human common sense before show the results.
Consistent refinement based on up-to-date information and guidance from stakeholders.
Saving Changes...
Anonymous
When using AI systems, what are some best practices for ensuring the results you receive are accurate, relevant, and aligned with your original goals? It is the same when reviewing the output of anything—does it answer the question posed, provide accurate information that can be sourced, and does it address the need given the context of the project. Saving Changes...
When using AI systems, what are some best practices for ensuring the results you receive are accurate, relevant, and aligned with your original goals?
Best practices for ensuring the results you receive are accurate, relevant and aligned with your original goal, are ensuring specificity is included in your prompt.
Using the CREATE (Character, Request, Examples, Adjustment, Type, Evaluation) prompt engineering method is a great starting point.
Adding organization specific as well as technical compatibility requirements helps ensure that desired results are in alignment with expectations. Saving Changes...
Anonymous
To get the most out of AI, it helps to treat it less like a search engine and more like a highly capable but literal-minded intern. Because AI models are trained on patterns, they are prone to "hallucinating" (making things up) or being overly vague if they aren't given clear guardrails.
Saving Changes...
Anonymous
To ensure AI outputs are accurate and aligned with your goals, you should treat the prompt as a structured specification rather than a casual question. Best practices center on the CLEAR framework: be Concise in your language, Logical in your task ordering, Explicit about constraints, Adaptive through iterative refinement, and Reflective by asking the AI to explain its reasoning (Chain-of-Thought). Specifically, you should provide a clear persona (e.g., "Act as a senior data analyst"), ground the request in verified context or uploaded documents to reduce hallucinations, and use "few-shot" prompting by providing examples of the exact format and tone you expect. Finally, always implement a human-in-the-loop verification step to cross-check factual claims against authoritative sources, as AI excels at fluency but can occasionally prioritize plausibility over truth.
PMI Team Member
