ticket-triage
bilingual-ticket-triage-dataset
bilingual-ticket-triage-dataset
Fully synthetic support-ticket dataset used to fine-tune bilingual-ticket-triage-adapter (QLoRA, Qwen2.5-3B-Instruct). Tickets are written in Roman Urdu, Urdu script, English, and code-mixed variants, mirroring how Pakistani customers actually write support emails.
No real customer data. All personas, names, addresses, and order numbers are fabricated. No real email addresses appear in the source seeds.
Content
Each record:
{… See the full description on the dataset page: https://huggingface.co/datasets/abuzarkhan/bilingual-ticket-triage-dataset.ghfht9074_582a26b2_support_ticket_triage
Nova Dataset: Support Ticket Triage
Overview
This dataset is a curated collection of customer support ticket messages labeled by priority and routing category. It is used to train triage models that automatically route incoming support requests to the appropriate team.
Description
The corpus was assembled by the Nova Customer Experience team and transformed from an upstream question-answering corpus. Each record contains a ticket subject, message… See the full description on the dataset page: https://huggingface.co/datasets/TianfuXinqu/ghfht9074_582a26b2_support_ticket_triage.ticket-triage-swbza19i
Support Ticket Triage Dataset
This dataset contains a set of customer support tickets that must be triaged
according to a prioritization policy.
Purpose
An analyst must read the prioritization policy and apply it to every open
ticket, deciding whether each ticket should be escalated or resolved
based on the rules in the policy. The final decision for each ticket is then
recorded in a structured report.
Layout
policy/prioritization_policy.json - the… See the full description on the dataset page: https://huggingface.co/datasets/Roy229/ticket-triage-swbza19i.