Hello, All
Newbie’s here.
When building a prompt for your knowledge base from your uploaded content, what are some of the best practices when creating the prompt? A few questions about the PDF formatting, the pdf I’ve loaded has pictures does the up format for those? I’m looking to maybe incorporating - picture recognition, - e commerce- maybe some other ideas, into my studio. I’m not sure if these things are feasible yet.
Similar question, but I want the AI to reference the specific images in the PDF and not “generate” ones. Also, I want it to reference the page and section (i.e. Chapter 7, 1-2.a1) when answering questions so users and go back and re-read any specifics.
The best practices for prompts depends on the use case. This video walks you through making our the FAQ bot for Pickaxe.
If you upload a PDF into the Knowledge Base, images will not be recognized. When you upload a PDF, the system scrapes all the text from it and stores them into the Knowledge Base.
Your Pickaxe can, however, recognize images that the end-user uploads. Here is a video on building chatbots with image recognition. But the Pickaxe will only recognize the images if they are image files. If the end-user uploads a PDF with images… same story. It will not recognize them.