Project Management Central
|
Please login or join to subscribe to this thread
|
|||
|
|||
To ensure the results you receive from AI systems are accurate, relevant, and aligned with your original goals, follow these three simple steps:
1. Design Expected Prompts with AI Assistance: - Craft clear and specific prompts to guide the AI in generating the desired output. - Use AI tools and suggestions to refine your prompts, ensuring they are easy to understand and aligned with your goals. 2. Evaluate the Results with AI Assistance: - Review the AI-generated results to check for accuracy and relevance. - Use AI tools to analyze and validate the responses, ensuring they meet your expectations and requirements. 3. Adjust Based on Your Knowledge: - Make adjustments to the prompts or results based on your expertise and knowledge. - Iteratively refine the prompts and review the outputs until they align perfectly with your original goals. By following these steps, you can effectively leverage AI systems to achieve accurate, relevant, and goal-aligned results. 
Two main steps are most useful at this moment: 1) use a well-defined prompt, with a detailed context and clear request, 2) check every output and do not take anything for gold.
I could not agree less with some of my great colleagues here. The quality of input as per the details prompts to Gen AI determines the expected output (Goal). One caveat though is that one needs to pay attention to not giving out sensitive information in a bid to providing more details information while providing the prompts.
AI is a powerful tool, but it's most effective when used in conjunction with human expertise. I encourage collaboration between AI systems and construction professionals to leverage the strengths of both.
Set Clear Objectives: Define what you want to achieve with the AI to guide its responses and actions effectively. Craft Specific Prompts: Use detailed and precise prompts to minimize ambiguity and direct the AI towards relevant outputs. Provide Continuous Feedback: Regularly review and give feedback on the AI's outputs to refine its performance and improve accuracy. Cross-Verification: Validate the AI's results against trusted sources or expert opinions to ensure their reliability. Contextual Awareness: Supply background information and contextual details to help the AI understand the task better. Regular Updates and Training: Keep the AI system updated with the latest data and retrain it periodically to maintain its relevance and accuracy. Ethical Considerations: Address potential ethical issues such as bias and data privacy to maintain trust and compliance. Expert Collaboration: Involve domain experts to verify the AI's outputs and ensure alignment with industry standards and practices. Monitor and Evaluate Performance: Continuously assess the AI's performance to identify areas for improvement and ensure it stays aligned with your goals. Document Changes and Learnings: Maintain records of updates, feedback, and changes to track the AI's development and ensure consistency with your objectives. 
Giovanni Alonso Alvarado Morales
Project Management/ Business Intelligence/ Strategy| @ RACSA
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.
Firstly, it is essential that a user is clear about the original goals as it may be quite easy to get carried away with responses.
The user should be able to identify responses that align with original goals. This means that the prompts should be well structured and in line with the original goals. 
Using AI requires patience and continuous learning, both for the user and the LLM. Unless working within an organizations tailored-specific LLM, users have to scrutinize the outputs against their organization's quality measures and against the user's expectations for the output.
When using AI systems, start with clear objectives and provide accurate, relevant input data. Verify sources, iterate and refine queries, use human oversight, and continuously test AI outputs against benchmarks to ensure accuracy and relevance.
I usually start with the RTF prompts just to ascertain the accuracy of the initial response. Then I like to add details and information to refine the response and see if the response is still reasonable. Lastly, I like to add question that go beyond the limits of my understanding and see if the results still seem reasonable.
|
|||
PMI Team Member
Please login or join to reply
|
'Human existence must be a kind of error. It may be said of it: "It is bad today and every day it will get worse, until the worst of all happens."' - Arthur Schopenhauer |
