AI Agents in Automated Workflow
An Experiment with an Educational Representative and a RAG Legal Advisor
YUSHIEN, 2025
Abstract
An Experiment with an Educational Representative and a RAG Legal Advisor
YUSHIEN, 2025
This paper explores the methods and applications of building AI agent prototypes using existing automation and speech synthesis tools such as n8n, Vapi, ElevenLabs, and Make. Two experiments are presented to demonstrate the workflow: The first case involves creating a simulated representative for OCAD University's Digital Futures program, capable of answering questions about application processes and program details, as well as offering consultation scheduling services. The second case focuses on developing a legal advisor agent using Retrieval-Augmented Generation (RAG) technology, integrated with a legal database to provide accurate guidance on visa-related regulations. This paper details the system design, technical integration, and prototype implementation processes, and discusses the experimental results along with future improvement directions.
Keywords: AI automation, AI agent, retrieval-augmented generation (RAG), multimodal interaction, API integration