Diseño e implementación de un prototipo funcional de agente inteligente para la automatización de contenido legal en instagram: caso Lawem abogados

Abstract

This degree Project describes the design and implementation of a functional intelligent agent prototype aimed at automating legal content for the Lawem Abogados law firm. The project addresses the need to optimize social media management, specifically Instagram, through the use of artificial intelligence. The methodology used was based on design thinking, allowing an efficient transition from problem identification to technical prototyping. The system architecture integrates cutting-edge tools such as Next.js for the interface, N8N for agent orchestration, and the Qwen 2.5 language model. A critical aspect of the development was the technological migration from Gemini, the initially planned model, to an AI model orchestrator called Ollama. This change was driven by the increase in Gemini’s costs and its requirement for a 50-dollar prepaid balance, which compromised the financial sustainability of the project. By implementing Ollama, AI inference was performed locally, eliminating API costs and ensuring data privacy. Due to the high impact and technical complexity of the system, a manual was developed as an annex document, detailing the step-bystep configuration and implementation processes. The results demonstrate that the agent not only reduces production times but also ensures technically and legally valid communication.