A Virtual Assistant for Child Protection Information Management System (CPIMS)

CPIMS Chatbot

Abstract

The Child Protection Information Management System (CPIMS) serves as an indispensable framework in managing child protection cases. To address the challenge of high volumes of repetitive inquiries at the CPIMS service desk, a virtual assistant was innovatively crafted, utilizing advanced Data Science and Artificial Intelligence to improve user experience. The CPIMS chatbot was meticulously conceived through a thorough user engagement process to identify key services and user demands. Training data, gathered from regional WhatsApp chats and CPIMS's own knowledge base, was instrumental in the chatbot's development. An Exploratory Data Analysis (EDA) was pivotal in discerning user intentions, shaping the subsequent data cleaning, feature engineering, and model construction phases. At its core, the virtual assistant is powered by a supervised machine learning model, which has recorded an accuracy rate of approximately 89.02%, illustrating its adeptness in accurately fielding user queries. Since its launch, the chatbot has serviced over 113 users, significantly alleviating the load on the service desk. The chatbot's integration, facilitated by Flask, provides a seamless and interactive user interface. Moving forward, the project will focus on expanding the chatbot’s functionalities, improving its precision, and tailoring it to user feedback. This initiative is not only a testament to the immediate improvements to CPIMS service desk operations but also a harbinger of the potential for advanced AI-driven innovations in child protection.