<![CDATA[This study presents the implementation of a hybrid chatbot designed to enhance the efficiency of information services on the PT Jasamarga Pandaan Toll website. Conventional rule-based chatbots often struggle to provide accurate responses when users employ varied sentence structures or synonyms. To address this limitation, a hybrid approach is proposed that combines keyword matching with Natural Language Processing (NLP) capabilities through the Dialogflow platform. The research employed the waterfall development model, consisting of requirement analysis, system design, implementation, testing, and maintenance stages. Data were collected through observation and interviews within the company, and then transformed into intents, entities, and training phrases. The keyword matching component was implemented using a structured database, while Dialogflow was integrated with a webhook to dynamically process toll rate queries. Black-box testing was used to evaluate functional accuracy, and usability testing was conducted to measure user satisfaction and response efficiency. The results demonstrate that the developed chatbot can accurately recognize user queries, respond faster, and provide more relevant answers compared to a rule-based approach alone. This implementation highlights the potential of hybrid chatbot technology to improve customer interaction, reduce reliance on manual service, and support digital transformation in the toll road industry.]]>
Copyrights © 2025
