Developing a GPT Assistant for Appmetrica Analytics System

In the context of rapidly advancing digitalization and exponentially increasing data volumes, analytics systems occupy a key position as essential tools for data analysis and visualization, and decision support. AppMetrica, which is a comprehensive toolkit for monitoring and analyzing mobile applications, is no exception. Its success in the market is closely tied to several fundamental factors, such as the user-friendly interface, the speed of query processing, and the quality of the analytics provided. Against this backdrop, the development of an AI-enhanced assistant represents a forward-looking approach aimed at augmenting the user’s interaction with the system. This graduation thesis is devoted to the process of creating an AI-assistant based on the YaGPT model, capable of answering users’ analytical questions. The assistant’s operating principle is based on dividing the tasks of data collection and data analysis between two agents. A crucial stage in the development will be training the agents through prompt generation and configuring their interaction.