Skip to content
Karmyug

Agent and customer chat

Support chat across industries with intent, entity and resolution labels.

20 languagesCustomer support

Overview

Written support conversations covering retail, banking, telecom, travel, healthcare and logistics queues. Turns carry intent, entities and a resolution outcome so the set supports routing and assistant training in one pass.

Specification

Format
JSONL
Labels
Intent, entity, resolution

Sample

Illustrative sample. It shows the structure, fields and metadata of this dataset with invented values, not real recordings or records. A real sample pack is sent on request.

Sample records for Agent and customer chat
idlanguageturnsdomainspeaker_rolesexcerptlabelreviewer_agreementqa_status
tx_0001English6Customer supportagent, customerCustomer reports a delayed refund and asks for a status update.resolvedagreeaccepted
tx_0002English12Customer supportagent, customerAgent confirms identity and explains what happens next.escalatedagreeaccepted
tx_0003English4Customer supportagent, customerUser asks a follow-up that depends on the previous turn.resolvedsplit, adjudicatedaccepted
tx_0004English9Customer supportagent, customerCustomer switches topic and the agent re-confirms the goal.clarifiedagreeaccepted
tx_0005English7Customer supportagent, customerConversation closes with a summary and a confirmation.resolvedagreeaccepted

Quality assurance

  • Taxonomy consistency audit
  • Double labelling on ambiguous intents
  • Agreement reported per intent

Want a real sample before you commit?

The sample above shows the structure. We also send a representative slice that includes the hard cases, plus the QA report and licence terms. No call required.