AI-based Text Classification for Support Triage
I built a support triage service using the Anthropic API to classify inbound tickets according to a client knowledge base before routing them. This system reduced first-response times and lowered the rate of ticket escalation to engineering. My work involved designing the pipeline, integrating the model, and iteratively refining ticket classification outcomes. • Designed, implemented, and monitored an LLM-based classification system for customer support tickets. • Labeled text data through AI-powered workflows to automate and enhance support issue routing. • Utilized Anthropic API and internal tools for model integration and iterative system tuning. • Mapped knowledge base categories to ticket data to inform and refine classification criteria.