All case studies
Field service
Field-service AI dispatcher
Tickets triaged, routed and booked without a human dispatcher in the middle.
Client: ServiceGrid
n8nGemini 3 FlashWhatsApp Cloud APIGoogle CalendarPostgresPagerDuty

The challenge
Every ticket at ServiceGrid needed a dispatcher to read it, decide how urgent it was, pick a technician, and confirm a time. That was the whole ceiling on same-day work: about 40 tickets a day, no matter what.
Solution design
Now a small team of agents reads the ticket, works out urgency and skill needed, checks the live calendar for a technician who fits, and offers the customer a slot over WhatsApp. Humans only get pulled in for edge cases: VIP customers, cross-region jobs, anything weird.
Architecture
flowchart LR
A[Ticket In] --> B[Triage Agent]
B --> C[Skill Matcher]
C --> D[Calendar Availability]
D --> E[Slot Proposer]
E --> F[WhatsApp Confirm]
F --> G{Accepted?}
G -->|yes| H[Book + Notify Tech]
G -->|no| I[Re-propose]Rendered as a Mermaid flowchart in the deep-dive report.
Workflow screenshots
Triage agent trace
Triage agent trace
Tool calls: classify_urgency → match_skill → check_calendar → propose_slot.
Dispatcher exceptions queue
Dispatcher exceptions queue
Only VIP + cross-region tickets hit the human queue.