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?
I don't disagree with the answers above, but I keep it very simple. Make sure your data is clean, ask specific questions, and review the outcome. All of this will depend on the AI tools you are using and your needs for using them. Once you have this figured out, you will be good to go.
Continuing review and improvement are essential in this case.
I hope that helps.
Regards,
I agree keeping it simple with some clear built in validation criteria to ensure it doesn't AI hallucinate and also reiterate and refine the prompt so that the AI learns how best to respond
Saving Changes...
James PriceProject Manager / SCRUM Coach| American Cancer SocietyAtlanta, Ga, United States
Great question! Not to be cycinical, a calculator will not aid a person that cannot read or count. One should have an idea of the expected range of outcomes before relying on GenAI or any tool. Often, I use GenAI to create a cost model using MS Excel. However, I must have knowledge & experience to verify correct formulas have been applied given model's underlying assumptions. One other method is accountability....any PM working on SOX project is familiar with completing endless questionaires to confirm that project's controls, processes and dataI are accurate and effective. Saving Changes...
It will always be necessary to refine and structure the prompts and rewrite them using best practices until the best answer that fits the project goals is obtained.
It's important to relate the context for which we are requesting an output. Going through and validate the output gives more insight on what we want and what to explore to further enhance the output. Adding more context and cycling through validation in my view helps to improve the output in an adaptable manner. Saving Changes...
Anonymous
I confess, i don't have lot of experience with AI, but i think that you have to be more accurate in demand and using good prompt formulas etc.
Na minha visão, o uso de padrões avançados de prompting é essencial para garantir que as respostas da IA permaneçam alinhadas aos objetivos do projeto, especialmente em ambientes dinâmicos. O padrão ReAct é particularmente relevante, pois combina raciocínio com ação e permite ajustes contínuos com base em feedback e mudanças no contexto. Isso ajuda o gerente de projetos a manter a coerência estratégica, adaptar decisões rapidamente e garantir que a IA apoie efetivamente a evolução dos objetivos do projeto ao longo do tempo. Saving Changes...
Ask for the sources and possibly ask the same question in several different ways to see if the AI comes back with the same or similar answer. If you're asking the same question several different ways and you're getting several different, conflicting answers, then that's a red flag that the AI's accuracy isn't at the level you need, which means you may need to teach the AI some more, and/or refine your prompt.
I prefer to use AI in areas where I am familiar with other sources of authority and personal experience. That allows me to audit and review responses with some degree of independence, independent of the output.
I ensure that I use AI in areas where I am familiar with other sources of authority and personal experience. That allows me to audit and review responses with some degree of independence, independent of the output. Saving Changes...
"People are always blaming their circumstances for what they are. I don't believe in circumstances. The people who get on in the world are the people who get up and look for the circumstances they want and, if they can't find them, make them."
PMI Team Member
