1. Background.

Universidad de San Andrés (UdeSA) wants to leverage AI Chatbot technology to provide students with a personalized conversational experience resembling a student interaction with a professor. 

2. Proof of Concept idea.

UdeSA provided Rubik with both video and text materials to implement:

A) A speaker-diarisation model that detects the speaker and transcripts the text so the AI Chatbot (Agent) detects the professor’s voice as a source of truth for the subsequent response. 

B) A virtual agent that acts as a source of truth for student interactions. 

3. Test scenario. 

The testers will try two agents (Agent 1 and Agent 2), one that includes both text-based documents and transcripted videos without knowing what agent are they interacting with. 

The testers will provide feedback based on their interactions with Agent 1 and Agent 2. 

Important ⚠

After every response, the agent will ask for feedback. Please provide the feedback, in case you don´t the agent will stop interacting, the feedback is needed to continue the conversation.