18h/week recovered
Replaced 6 weekly manual reports with automated dashboards and email digests. Team saved 18 hours per week.
n8nGA4Google AdsResend
A SaaS company processing 400+ support tickets per week was struggling with routing. Every ticket landed in a single queue, and a support lead spent 2–3 hours daily manually reading, categorising, and assigning tickets to the right team member. Mis-routes were common — roughly 25% of tickets went to the wrong person first — which added a full day to average resolution time.
We proposed an AI-powered triage layer that would sit between ticket creation and assignment. The model would read each ticket, classify its category and urgency, extract key metadata (affected product, customer tier, sentiment), and route it to the best-matched agent — all within seconds of submission. The goal was 80% automation with full human fallback for edge cases.
We built the triage system using n8n for orchestration and OpenAI’s API for classification. When a ticket is created in Intercom, n8n triggers a workflow that sends the ticket content to a fine-tuned prompt. The model returns a structured response: category, urgency level, affected product, and recommended assignee based on current team capacity and expertise. High-confidence classifications (80% of tickets) are auto-routed. Low-confidence tickets are flagged for human review with the model’s suggested classification pre-filled.
The system auto-triages 80% of incoming tickets with 95% accuracy. Average resolution time dropped by 55% because tickets reach the right person immediately instead of bouncing between queues. The support lead recovered 2–3 hours per day from manual triage. Customer satisfaction scores improved by 18 points, largely driven by faster initial response times.
Replaced 6 weekly manual reports with automated dashboards and email digests. Team saved 18 hours per week.
Rebuilt the onboarding flow for a B2B SaaS platform. Trial-to-paid conversion increased from 8% to 19%.
Replaced a manual lead qualification process. Time to first contact dropped from 4 hours to 4 minutes.
