AI Ticket Classification & SLA Escalation System
AI & Machine Learning
Designed and built an AI-powered ticket processing and SLA monitoring system for an online bookstore to automate email classification, ticket routing, and escalation workflows.
Problem: Customer support teams manually processed incoming emails, resulting in inconsistent ticket prioritization, delayed responses to critical requests, no standardized SLA management, and limited visibility into support operations.
Solution: Developed an end-to-end automation using n8n, OpenAI GPT, Gmail API, Google Sheets, Telegram Bot API, and LangChain Structured Output Parser.
The workflow automatically retrieves incoming emails, normalizes message data, and uses AI to classify each ticket by category, priority (P1–P4), responsible team, customer sentiment, and SLA requirements. Structured outputs ensure reliable JSON responses for downstream automation.
Spam and irrelevant emails are filtered automatically, while valid tickets are logged into a centralized Google Sheets database and routed to the appropriate support owner with instant Telegram notifications.
A separate SLA monitoring workflow runs every five minutes, checking unresolved tickets for SLA breaches and triggering multi-level escalations—from owner reminders to team lead notifications and management alerts—ensuring critical requests are never overlooked.
The project included prompt engineering, structured AI outputs, automated ticket routing, SLA policy implementation, escalation logic, and scalable workflow design within n8n.
Results:
Reduced first response time from 30–60 minutes to under 1 minute
Automated classification of 100% of incoming support tickets
Eliminated missed critical (P1) requests through automated SLA monitoring
Centralized ticket logging and reporting for complete operational visibility
Saved 50+ support hours per month
Scalable support workflow without increasing headcount
Problem: Customer support teams manually processed incoming emails, resulting in inconsistent ticket prioritization, delayed responses to critical requests, no standardized SLA management, and limited visibility into support operations.
Solution: Developed an end-to-end automation using n8n, OpenAI GPT, Gmail API, Google Sheets, Telegram Bot API, and LangChain Structured Output Parser.
The workflow automatically retrieves incoming emails, normalizes message data, and uses AI to classify each ticket by category, priority (P1–P4), responsible team, customer sentiment, and SLA requirements. Structured outputs ensure reliable JSON responses for downstream automation.
Spam and irrelevant emails are filtered automatically, while valid tickets are logged into a centralized Google Sheets database and routed to the appropriate support owner with instant Telegram notifications.
A separate SLA monitoring workflow runs every five minutes, checking unresolved tickets for SLA breaches and triggering multi-level escalations—from owner reminders to team lead notifications and management alerts—ensuring critical requests are never overlooked.
The project included prompt engineering, structured AI outputs, automated ticket routing, SLA policy implementation, escalation logic, and scalable workflow design within n8n.
Results:
Reduced first response time from 30–60 minutes to under 1 minute
Automated classification of 100% of incoming support tickets
Eliminated missed critical (P1) requests through automated SLA monitoring
Centralized ticket logging and reporting for complete operational visibility
Saved 50+ support hours per month
Scalable support workflow without increasing headcount