Planly: an AI-Driven Assistant for Communication and Productivity Tasks
Course Instructor
Pramod Gupta
Abstract
Modern students often rely on multiple disconnected applications to manage tasks, schedules, and academic responsibilities, resulting in fragmented workflows and increased cognitive overhead. This project presents an AI-assisted productivity application designed to unify task management and calendar scheduling within a single conversational interface.
The system integrates a locally hosted Large Language Model (LLM) with a mobile application built using Flutter as the main app engine, enabling users to interact with it through natural language. By leveraging real-time AI response streaming and custom in-app action parsing, the application interprets user input and converts it into structured operation calls such as creating to-do lists, adding tasks, and scheduling calendar events for app functionalities, all in a modern user interface.
Key features include dynamic conversation handling, persistent data storage, and a confirmation-based action execution system to ensure accuracy and prevent model hallucination or other unintended operations. The application demonstrates how natural language interfaces can simplify user interaction and reduce the complexity associated with traditional productivity tools.
This work exemplifies the potential of integrating AI-driven interaction into everyday applications, providing a more intuitive and efficient approach to managing academic workflows.
Planly: an AI-Driven Assistant for Communication and Productivity Tasks
Modern students often rely on multiple disconnected applications to manage tasks, schedules, and academic responsibilities, resulting in fragmented workflows and increased cognitive overhead. This project presents an AI-assisted productivity application designed to unify task management and calendar scheduling within a single conversational interface.
The system integrates a locally hosted Large Language Model (LLM) with a mobile application built using Flutter as the main app engine, enabling users to interact with it through natural language. By leveraging real-time AI response streaming and custom in-app action parsing, the application interprets user input and converts it into structured operation calls such as creating to-do lists, adding tasks, and scheduling calendar events for app functionalities, all in a modern user interface.
Key features include dynamic conversation handling, persistent data storage, and a confirmation-based action execution system to ensure accuracy and prevent model hallucination or other unintended operations. The application demonstrates how natural language interfaces can simplify user interaction and reduce the complexity associated with traditional productivity tools.
This work exemplifies the potential of integrating AI-driven interaction into everyday applications, providing a more intuitive and efficient approach to managing academic workflows.
