See full schedule

Most business knowledge is stuck in PDFs, slide decks, images and reports but text only agents miss almost all of it. We'll fix that live, parse documents with Jina visual language model, assemble context in Elasticsearch, and run one agent task twice, once on raw text then with context. You’ll see accuracy, latency and token counts side by side. We’ll dig into the gotchas building visual context and share a repo you can try for yourself.
This is a live workshop hosted by Elastic. Please bring your laptop or device if you'd like to participate. 

The knowledge your agents can't see: Building visual context

Speakers