Skip to main content

Announcement: Triage 2.0 is in Limited Preview

  • February 20, 2025
  • 3 replies
  • 331 views

andrew.mairena
Forum|alt.badge.img

For all existing customers with Triage enabled, we are excited to announce Triage 2.0! Triage 2.0 is a significant upgrade to our existing Triage product. The product automatically classifies newly created tickets from digital channels such as Chat, Email and Portal.

What are the advantages of this new architecture?

  1. Relaxed data requirements: Less data and no data distribution requirements required for Triage, due to the embeddings + index infrastructure.
  2. Transparent model explainability: We can inspect all the nearest neighbors to see clearly why a prediction was made on a specific ticket for troubleshooting purposes.
  3. Improved overall model performance: Based on offline analysis of existing set of customers, Triage 2.0 performs as good or better than the current model architecture, irrespective of the digital channel, e.g. chat, email, portal.

For a deeper technology overview, including detailed architecture diagrams, consult our documentation.

Do you want to self-nominate for Limited Preview?

Visit and read the form here

For additional questions, please ask in thread. 

3 replies

andrew.mairena
Forum|alt.badge.img
  • Author
  • Community Manager
  • March 4, 2025

Currently, Triage 2.0 is only supported in US Data Center. The next available Data Center, EU, will be ready by March 31. 


priya.kornalius
Forum|alt.badge.img+2

What are the differences between the earlier Triage version and 2.0? How is the experience changing between these 2 versions?


afleury
Forum|alt.badge.img+3
  • Inspiring
  • March 24, 2025

Getting an access forbidden error message when I try to click on the link https://community.moveworks.com/p/triage-2.0 to self-nominate for the limited preview.


andrew.mairena
Forum|alt.badge.img
  • Author
  • Community Manager
  • March 26, 2025

Triage 1.0 vs 2.0:

  • What’s the same? Key configuration controls - the Triage ticket filters, heuristics rules, and metadata cache system (for ServiceNOW) - and the functionality of intercepting and making predictions on newly created tickets. This is all the same between both versions.
  • What’s different? The underlying technology evolved from distilbert, a deterministic model, to MPNET, an embeddings model. Basically turning the classification task into a search problem, e.g. Embeddings Nearest Neighbor Search. This transition has created 3 key benefits aforementioned in my original post -  Relaxed data requirements, Transparent model explainability, & Improved overall model performance.