This paper presents the development and functional analysis of an application designed to optimize corporate communication through automatic email classification and response generation using the Google Gemini API. The study addresses the challenge of increasing message volumes in organizations, which negatively impacts productivity and response times. The adopted methodology was experimental, integrating a Python backend architecture with strategies to mitigate rate limitations. Tests were conducted using a dataset of 1000 messages to evaluate response times and classification accuracy between productive and unproductive emails. The results demonstrate that the solution provides efficiency gains in sorting and response standardization, validating the use of advanced language models in corporate environments, despite computational cost limitations at large scales.