Use cases · headline

Ticket and inbox triage: one call instead of five models

Route, score urgency, read the customer’s temperature, extract the order ID. One request over the raw email returns a filled triage card, each field with its probability.

The usual stack

Five services, five calls, five latencies added up: an intent classifier, a sentiment model, a moderation model, an extractor and a router. Each one reads the same email again.

One Simo request

Which team handles this?          →  billing: 0.91, shipping: 0.05, technical: 0.04
How urgent is it? (0–3)           →  2.1
How frustrated is the customer?   →  2.4 of 3
What is the order ID?             →  ["A-20417"]    (preview)

Add a sixth question and the response time barely moves. Reading is the cost; questions are nearly free.

What you do with it

Route automatically when the team probability is high; send ambiguous tickets to a human triager.
Escalate on urgency and frustration regardless of team.
Attach the extracted order ID so the agent opens the ticket with the context already filled in.

Which model

Simo-1. Order-ID extraction is in preview. On the Banking77 intent-routing benchmark, Simo-1 Pro scores 81.8% (see accuracy).

Frequently asked questions

Can Simo replace my intent classifier and sentiment model?

Yes: team routing, urgency, sentiment and extraction are all questions to the same model over the same read of the email.