INAI | Devpost

Inspiration

The inspiration for INAI came from observing how tech teams spend countless hours in discussions where important action items are often lost. I wanted to build a system that bridges the gap between team communication and task management, making sure no task is missed and workloads are distributed fairly. The vision was simple: turn team chats into structured tasks and balanced workloads automatically.

What I Learned

How to design and build a custom messaging system tailored for structured conversations.

Applying AI and NLP techniques to extract action items from noisy, unstructured chat messages.

Building a React dashboard with charts to visualize real-time workloads and task distribution.

Managing real-time synchronization between conversations, tasks, and workload balancing.

How I Built INAI

Messaging System – Designed and implemented a custom communication layer where users can send and organize messages.

AI Task Extraction – Used an LLM to summarize conversations and detect action items, outputting them as structured JSON.

Task Management Layer – Converted detected action items into trackable tasks within INAI’s system.

Workload Equalizer – Developed an algorithm to calculate workload scores and recommend fair task distribution.

Dashboard – Built a React-based dashboard with dynamic visualizations (Recharts) showing team workload and new tasks in real time.

Challenges I Faced

Messy conversations: Separating casual chatter from actual tasks was difficult for the AI.

Workload balancing logic: Getting the equalizer right required multiple iterations to factor in both availability and complexity.

Real-time sync: Ensuring that new tasks and completed tasks updated instantly across the system.

Scalability: Designing a messaging system that could handle multiple teams and conversations simultaneously.

Time constraints: Delivering a polished, functional product while managing deadlines.

With INAI, I learned how to merge AI, messaging, and workload optimization into one powerful productivity tool. The project not only solved a real-world problem but also taught me how to design AI systems that handle messy, dynamic team data while staying practical and user-friendly.