Abstract:
In the dynamic landscape of call centers, ensuring the quality of interactions between agents and customers is necessary. This project introduces an innovative AI-based tool designed to streamline the analysis of rejected or flagged calls within the Quality Assurance (QA) department. The primary goal is to empower QA analysts by expediting the identification and examination of critical moments during calls, particularly focusing on negative emotions. The target audience for this tool comprises QA analysts working in call centers, providing them with a comprehensive dashboard equipped with statistical visualizations and user-friendly
functionalities. Additionally, administrators have access to an interface for managing analysts efficiently. Key features include the ability for analysts to upload audio recordings for analysis, view call history, and make use of graphical representations of emotional patterns. The administrators can oversee analyst performance, manage user accounts, and access comprehensive stats. The system is designed using the Material UI Tailwind Dashboard, ensuring an intuitive and visually appealing user experience. Utilizing a technology stack consisting of React for the frontend, MongoDB for the database, NodeJS and ExpressJS for the backend, and Python with Flask for emotion recognition models. The iterative development process allows for continuous improvements and adjustments based on feedback gathered from the FYP panel members. With a focus on transactional user journeys, the system is optimized for responsive use on larger screens, primarily in the Chrome browser, and can adapt to smaller screens such as iPads. The website's maintenance and updates will be managed by the project team post-launch, with regular feedback loops ensuring ongoing enhancements. This AI-driven QA analysis tool is poised to significantly enhance the efficiency and effectiveness of call center quality assurance processes.
