Desarrollo de un Agente Conversacional basado en Inteligencia Artificial, para Solucionar Fallas de Hardware de una Computadora.

Abstract

Abstract: This project develops a Conversational Agent based on Artificial Intelligence to help non technical users resolve hardware failures in their computers. The solution combines a locally deployed large language model (in this case, Mistral 7B) with a Retrieval-Augmented Generation (RAG) architecture that integrates multiple services, orchestrated by N8N, and a vector database (Chroma DB), enabling queries to a specialized knowledge base. The orchestration system in N8N is deployed in Docker containers and provides an intuitive web interface via Streamlit. The conversational agent interprets the user's problem description and retrieves information from documented solutions. This implementation demonstrates a low-cost, fully local alternative, reducing downtime and dependence on technical support personnel.