RAG, the proper way · AI Automation Society

RAG, the proper way

I just watched your video on RAG—great content!

I had a quick question/remark: shouldn’t different file types be treated differently during processing? Perhaps I missed it in the video, but I’m curious about your approach to handling this.

Here’s what I’m envisioning (though I haven’t tried it myself yet): a workflow where I can drop all my data into a folder, and the system processes it intelligently, identifying what’s been successfully ingested, what’s missing from the vector database, and any issues encountered. Additionally, how do you manage updates to existing files or handle deletions?

I’d really appreciate your thoughts or a deeper dive into proper file handling for these scenarios.

Thanks again for the insightful content!

Regards,

Jan

6 comments