Hey, thank you for watching! Hope this step-by-step walkthrough would help you understand the process of building an AI Agent that can execute tasks in real time. You can quickly reproduce the workflow using the resources I provided from the link in the description. How do you like this format? Please let me know in the comments - your feedback will help me shape future content!
I would love for you to see on creating using the same assistant with medical records and history of specific events throughout someone’s life because mine tends to hallucinate
What works for me to address hallucination is tons of trial and error in testing the combination of prompts and the AI model to use until I get the result I wanted. OpenAI 4o model works best for me when building AI Agent workflows or any tool calling implementations, while I use Claude for RAG projects. Anything that interacts with documents like retrieval, and when you want the responses to be more grounded with your specific data, Claude models seems to be more accurate for this types of use-case. In the end it all depends on how you structure your prompts, and test rigorously with your AI model of choice 🙂
@federicoarg00, the pleasure is mine! Glad that you liked this. Let me know if you have anything in mind around this topic that you want me to cover in my next videos.
Honestly I haven't tried it yet but I think I found some video showing that it's possible to implement multiple language. Not sure, I'll be checking on it, thanks man!