Work · Support tooling

Signal

Model-assisted triage that routes confident cases automatically and gets everything else to the right person faster.

Client

Confidential · B2B software

Sector

Support tooling

Duration

12 weeks

Our role

Applied ML, backend, product design

Triage · queue 42

Refund request0.94

routed → Billing

Contract question0.88

routed → Legal

Integration bug0.41

needs a human

54%

tickets auto-routed

−41%

first-response time

96%

routing precision on the eval set

The challenge

Nine support agents triaged every inbound ticket by hand. Around half were routine and repetitive; the rest needed a specialist who often saw them a day late.

Our approach

We built a retrieval-backed classifier with an explicit confidence threshold and an evaluation harness run on every change. Above threshold, tickets route and draft themselves. Below it, they go to a person with the retrieved context attached — the model never closes a ticket alone.

Deliverables

What we shipped

  1. 01Retrieval-backed triage and routing service
  2. 02Confidence thresholds with human-in-the-loop fallback
  3. 03Evaluation harness with a labelled regression set
  4. 04Agent-facing draft replies with cited sources
  5. 05Quality dashboard tracking drift and override rate

Technology

  • Python
  • Postgres
  • pgvector
  • Evals

Have something like this to build?

Start a project