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
Field-service AI dispatcher

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.