If you want to make an interactive application AI-enabled with a LLM, you will need more than just a chatbot interface.

In this makerspace, we will explore how LLMs can incorporate non-verbal user actions into their prompts without using a vector index.

We will look at how to leverage the natural entity IDs in your application (user_id, session_id, order_id, etc) to retrieve fresh information related to recent user activity as well as historical user activity (such as orders placed).  We will do this in the context of the H&M fashion recommender project, based on the open-source H&M dataset.