Agentic AI Framework for Operating System Automation
Specially designed for Windows users - Automate tasks without human intervention using advanced AI technology
AssistiveWorkmate (AW) is an innovative framework designed to automate and manage operating system (OS)-level tasks through an Agentic AI approach. Unlike conventional automation tools, AW functions as an intelligent agent that accepts natural language commands, interprets them, and executes corresponding actions directly within the OS environment.
Born from the frustration of repetitive troubleshooting and system recovery challenges faced by cybersecurity students, this project aims to provide a one-step solution for system management, enabling users to perform complex configurations, fix issues, and streamline workflows simply by interacting with an AI-powered assistant.
Converts natural language queries into executable automation scripts using Google's Gemini API
Direct interaction with Windows OS for real-time automation tasks including file management and process handling
Automated detection and resolution of common OS-level issues, reducing dependency on manual problem-solving
Simple text-based input mechanism for seamless interaction by both technical and non-technical users
Immediate task execution with transparent step-by-step automation process and progress tracking
Users can review automation steps before execution, maintaining full control over system operations
User provides their query in the chat field using natural language
System sends predefined prompt + user prompt to Gemini API
Gemini provides both user-friendly reply and automation JSON
User can review steps and choose whether to automate the task
System executes automation steps with real-time progress tracking
React + Next.js + Tailwind CSS
Google Gemini 2.0 Flash Model
Python + PyAutoGUI + System Commands
Windows 10 and above
CSE (Cybersecurity) • Team Member 1
Lead Developer & System Architecture
CSE (Cybersecurity) • Team Member 2
AI Integration & Testing
Institution: Madanapalle Institute of Technology & Science (MITS)
Developed under: NxtGenSec • Date: August 25, 2025
Expand to Linux and macOS for wider accessibility
P2 2026Speech-to-text functionality for hands-free operation
P3 2026Multiple AI agents working in parallel on different tasks
P4 2026Real-time monitoring with predictive analytics
P5 2027Microsoft MCP integration with RBAC and encryption
P6 2027Cross-device automation preferences and history
P7 2027