FURI | Spring 2025

Guardian Angel: AI-Driven Daily Life Planning System

This research presents an AI-driven framework, “Guardian Angel,” that utilizes Large Language Models (LLMs), specifically ChatGPT 4 and Gemini 1.5 Pro, integrated with various APIs to optimize real-world daily life planning. Unlike traditional algorithms, which are limited in scope, LLMs can dynamically generate personalized and adaptable plans. Initial groundwork has been laid, including a theoretical review of LLMs in planning applications and the identification of key APIs such as Google Maps and Google Flights for integration. A preliminary dataset has been created, focusing on cases like meeting/trip scheduling, restaurant recommendations, hiking assistance and human body cycle tracking. An early-stage prototype is in development using Python programming, testing basic API interactions for retrieving location-based data. The next phase involves refining dataset collection, enhancing model prompt design, and implementing API-driven planning logic. By combining AI’s reasoning capabilities with real-world data, this approach aims to create a scalable and intelligent planning assistant tailored for daily tasks.