01
Explainable triage is what the office trusted
Every priority came with a written reason the dispatcher could read. That single design choice turned skeptical office staff into the system's strongest advocates within the first weeks.
A voice and chat intake layer, an automated triage engine, and a dispatch integration that answers every inbound request, prioritizes it, and books it without a person in the loop.
The challenge
A growing field-services company was losing jobs it never knew existed. Calls rang out after hours, overflow calls during storms went to voicemail, and the ones that were answered landed on a dispatcher who had to figure out urgency, location, and technician availability by hand before anything was booked. Triage was the bottleneck: every request waited for a human to read it, classify it, and decide where it went.
The company needed every inbound request answered immediately, classified correctly, and routed to the right technician slot, with emergencies escalated to a person and routine jobs booked automatically. It also needed the office to trust the system, which meant every automated decision had to be explainable after the fact.
The solution
InsightSol built a three-layer architecture. The intake layer answers calls and chats around the clock with a voice agent on Twilio and a web chat agent, transcribes and structures each request, and captures photos and location where relevant. The triage engine classifies intent and urgency with a model constrained by written rules, enriches the request from the CRM and service history, and assigns a priority with an explanation attached.
The dispatch layer turns a triaged request into a booked job. It queries technician availability, skills, and location, proposes slots, confirms with the customer by voice or text, writes the job to the dispatch system, and notifies the technician. Emergencies route to an on-call human with full context in seconds. Everything else is booked, logged, and auditable without office staff touching it.
Headline result
42%
reduction in operational triage time
System architecture
Voice and chat intake
Twilio telephony with streaming speech recognition, web chat agent, photo and location capture
Triage engine
Intent and urgency classification under written rules, CRM enrichment, explainable priority assignment
Dispatch integration
Availability, skills, and location queries; slot proposal; confirmation; job creation; technician notification
Escalation path
Emergency detection routed to an on-call human with transcript and context attached
Orchestration and side effects
n8n workflows for notifications, CRM sync, and follow-up sequences
Telemetry
Per-request traces, triage accuracy tracking, and owner dashboard
Quantified results
01
Every priority came with a written reason the dispatcher could read. That single design choice turned skeptical office staff into the system's strongest advocates within the first weeks.
02
Half-automation would have kept a dispatcher watching the queue all day. Routing genuine emergencies to a person immediately is what made it safe to let routine jobs book themselves.
03
The 42 percent reduction is measured from the moment a call or chat arrives, not from when a human opened the request. Measuring from the customer's side is the only number that reflects their experience.
Disciplines involved
The audit produces the same diagrams and unit economics this system started with, for your workflow and your data.