First test is just for school
Regular AI Agents have 3 main steps:
- Task creator
- Executor
- Prioritizer
Since i’m trying to automate learn, my executor needs to be a browser agent as well.
knowledge base = ""
My browser agent should be simple: Takes in: Thing that it is trying to find Steps: gets website data to see if the information exists, if not, scan for all clickable elements and use AI to figure out what element will get closer to the information. If there is a downloadable pdf, then download and scan that for information too. Add any relevant information to the knowledge base.