ResolveIQ (SmartDesk / NH26)

AI-Powered Customer Support & Escalation Engine

React 19ViteTailwind CSS v4Socket.io-clientAxiosNode.jsExpress

National 2026 Hackathon (Surat)Real-time Socket.io ticket escalationTailwind CSS v4 & React 19 Stack

The problem

Built during my first out-of-state hackathon at Surat, Gujarat (National 2026 Hackathon, Team Spark-AI, submission 80-spark-ai). Customer support desks struggle with high ticket volumes, failing to classify issue severity or escalate high-risk tickets to human agents in real time.

We needed an instant-escalation workspace chat application that bridges automated AI query resolution and live human agent socket queues.

Key decision

We used React 19's virtual DOM paired with Tailwind v4's build-time engine for blazing fast UI renders. For real-time state syncing across users and support agents, we chose Socket.io to establish persistent, bi-directional TCP connections instead of polling.

How it works

The code is structured as an interactive SPA within submissions/80-spark-ai/:

Views: LandingPage.jsx (ResolveIQ onboarding), ChatPage.jsx (client chat interface), AgentLogin.jsx, and AgentDashboard.jsx (real-time ticket queue management).
Components: MessageBubble.jsx, TicketCard.jsx (metadata visualizations), SeverityBadge.jsx, and TypingIndicator.jsx.
State: Global user metadata handled via UserContext.jsx.

Results

Delivered a fully responsive support portal containing smart severity detection badges, typing indicators, and seamless Socket.io-driven handoffs from AI replies to the Agent dashboard ticket list.

What went wrong

Nothing broke permanently, but we ran out of coffee at 3 AM in Surat, and my buddy started naming variables in Gujarati. Clean code guidelines do not survive sleep deprivation.