An AI-powered assistant for automating customer support tickets, classifying issues, and suggesting resolutions.
AI Ticket Assistant is designed to revolutionize traditional customer support workflows. It reads incoming support tickets, classifies them using LLMs, and provides recommended solutions with high accuracy. It reduces agent workload by automating repetitive tasks and intelligently escalating complex issues. The system integrates with customer CRMs and enables contextual support through API-connected data streams.
I was responsible for integrating the AI model and designing the flow from ticket submission to resolution. This included setting up the LLM communication layer, defining fallback rules, and ensuring ticket metadata and history are preserved. I also implemented secured API routes, JWT-based role authentication, and an admin dashboard for real-time system monitoring.
MERN Stack, Inngest AI, Inngest Event-Driven Architecture, JWT and Auth.
As it was my first time dealing with an AI agent project, integrating AI posed unique challenges including handling unpredictability in responses, prompt design, and fallback systems. Ensuring consistent outputs from the AI, setting confidence thresholds for auto-responses, and debugging LLM behavior was also a learning curve.
Through this project, I gained hands-on experience in building production-ready AI workflows. I learned how to work with event-driven systems, structure scalable APIs, and manage role-based access control effectively. Understanding the importance of user feedback loops and how to improve AI reliability through metrics was a key takeaway.