AI INSIGHTS: UNVEILING UK ENERGY CONSUMPTION WITH LANGCHAIN- POWERED CHATBOTS

Abstract

Post-occupancy evaluation (POE) is crucial for assessing building performance, but traditional approaches falter under data volume and lack personalization. Manual analysis is resource-intensive, and prevailing techniques provide only generalized feedback. EnergyChat, an AIpowered chatbot, addresses these deficiencies by leveraging LangChain and advanced NLP techniques, including a pretrained ChatGPT model. Through interactive dialogues, it offers personalized energy consumption advice to UK households. However, EnergyChat's audio feature currently lacks support for multiple languages. Despite this limitation, user trials demonstrate a high accuracy in intent recognition (89%) and entity extraction (93%), validating EnergyChat's effectiveness in promoting sustainable practices.