Skip to content
Dataset By Humans
On requestConversations· Customer service

Customer Support Conversation Dataset (Vietnamese & English)

Multi-turn chats between customers and support agents for e-commerce, delivery, banking and telecom, written and role-played by trained people in natural Vietnamese and English.

On request: this dataset is produced to your specification after you order. Typically 2–3 weeks for 500 conversations and 5–6 weeks for 2,000.

Pricing

Fixed prices for standard packages, produced to your spec. Pick a package and a license to see the exact total, or ask for a custom quote.

Packages compared
FeatureStarterProCustom
conversations5002,000Any quantity
What's included
  • 500 conversations in JSONL
  • Intent and outcome labels
  • Datasheet, schema, license and checksums
  • 2,000 conversations in JSONL
  • Everything in Starter
  • Industries and intents chosen by you
  • Sentiment per turn
Your quantity, annotations and deadline
Datasheet, license, checksumsIncludedIncludedIncluded
Research license$590$1.18 / conversation$1,790$0.90 / conversationQuote
Commercial license$1,790$3.58 / conversation$5,500$2.75 / conversationQuote
Exclusive license$5,900$11.80 / conversation$17,900$8.95 / conversationQuote

How ordering works

  1. 1Choose a package and license and place your order. No payment is taken online.
  2. 2Within 48 hours our team contacts you to confirm the spec and invoices a 50% deposit.
  3. 3We produce and deliver in parts so you can review early. Typically 2–3 weeks for 500 conversations and 5–6 weeks for 2,000.
  4. 4Pay the balance and download the final dataset.

Custom quote

Need another quantity, an Enterprise license or annotations for this dataset? Available here:

  • Your policies and tone guide
  • Sentiment per turn
  • Escalation labels
  • Voice-call transcripts
Request a custom quote

Overview

Support chatbots in Southeast Asia must handle code-switching, teen-code abbreviations, angry customers and policies that differ by company. Real chat logs are full of personal data and rarely legal to train on.

Trained writers and former support agents role-play realistic conversations from scenarios we design with you: intents, customer personas, policies and edge cases. Each conversation is labelled with intent, outcome and quality, and contains only invented personal details.

Example records

Illustrative records in the delivered format. Request the sample pack for real records written by our experts.

  • Record 1 of 1 · illustrative

    {
      "id": "csc-000781",
      "intent": "order-not-received",
      "outcome": "resolved",
      "industry": "e-commerce",
      "language": "vi",
      "messages": [
        {
          "role": "customer",
          "text": "shop oi don hang DH-48213 cua minh 5 ngay roi chua thay giao"
        },
        {
          "role": "agent",
          "text": "Dạ em chào anh/chị ạ. Em kiểm tra thì đơn DH-48213 đang ở kho Bình Dương và bị trễ do lượng hàng tăng. Dự kiến giao trong 2 ngày tới ạ."
        },
        {
          "role": "customer",
          "text": "lâu vậy, mình cần gấp cho thứ 7"
        },
        {
          "role": "agent",
          "text": "Em rất tiếc vì sự chậm trễ. Em đã chuyển đơn sang ưu tiên giao hỏa tốc miễn phí, đơn sẽ đến trước thứ 6. Em gửi anh/chị mã giảm 30.000đ cho lần mua sau ạ."
        }
      ],
      "sentiment": [
        "negative",
        "neutral",
        "negative",
        "positive"
      ]
    }

Get a free sample pack

We email you real conversations from this dataset.

Used only to send your samples. Privacy policy.

Technical specifications

Authors
Trained writers and former customer support agents
Industries
E-commerce, delivery and logistics, banking and e-wallets, telecom, travel
Language
Vietnamese, English and natural code-switching
Fields
Messages (role, text), intent, sub-intents, sentiment, outcome, policy references, quality score
Length
Typically 4–16 turns per conversation
Format
JSONL (one record per line), UTF-8; CSV on request
AI use
Written by people; screened for LLM-generated text. Any AI assistance is disclosed in the datasheet
License
Research, Commercial, Enterprise or Exclusive

Use cases

  • Support chatbot fine-tuning

    Teach assistants your tone, policies and escalation rules with realistic multi-turn examples.

  • Intent and sentiment models

    Labelled intents, sentiment and outcomes for classifiers and routing.

  • Agent assist

    Train suggestion tools that draft replies for human agents.

  • Bot evaluation

    Use conversations as scripted test scenarios for your production bot.

How this data is made

  1. 1Scenarios designed with you: intents, personas and edge cases
  2. 2Conversations role-played or written by trained people
  3. 3A reviewer checks tone, accuracy and policy compliance
  4. 4Names, phone numbers and addresses are invented, never real
  5. 5Screened for copied text and LLM-generated writing
  6. 6Intent, outcome and quality labels added per conversation
  7. 7Exported as JSONL in a chat-message format
  8. 8Datasheet and license packaged with the delivery
Read our full process

Provenance & legal

  • 100% made by people: no scraping, no generative AI
  • Datasheet documenting how the data was made, checked and its limitations
  • Commercial license that lets you keep models trained on the data
  • Supports training-data documentation under the EU AI Act
Licensing options

Dataset-specific notes

  • Names, phone numbers, order IDs and addresses are invented; no real chat logs are used.
  • Your policies and tone guide can be applied under NDA, with the result licensed exclusively to you.
  • Difficult cases (angry customers, fraud attempts, out-of-scope requests) are included on purpose.

Delivery format

One JSON record per line, with the fields shown in the example records above and a SCHEMA.md describing each one.

Each delivery contains:

customer-support-conversations-v1/
├── data/conversations.jsonl
├── data/sample.jsonl
├── SCHEMA.md
├── DATASHEET.md
├── LICENSE.pdf
└── checksums.sha256

Frequently asked questions

Are these real customer chats?

No. They are role-played or written from scenarios, so they contain no real personal data and are safe to train on.

Can you follow our company's policies?

Yes. Share your policies and tone guide under NDA; those conversations are licensed exclusively to you.

Do you cover informal Vietnamese?

Yes. Customers write the way people really chat: no diacritics, abbreviations, emoji and mixed Vietnamese–English.

How does ordering and payment work?

Create a free account and place your order on this page; nothing is charged online. Our team contacts you to confirm the spec and invoices a 50% deposit, with the balance due on final delivery. The deposit is refunded in full if we cannot deliver the agreed spec.

What is the difference between non-exclusive and Exclusive?

With Research or Commercial we may license the conversations produced for your order to other buyers later. With Exclusive they are never licensed to anyone else or added to our catalog.

Can I order a different quantity or extra annotations?

Yes. Request a custom quote with the quantity, annotations and deadline you need. Prices per record fall as quantity grows.

Need a variation?

Different topics, languages, difficulty, quantity or labels? Describe it and we reply within 48 hours with matching samples and a quote.

Request a quote
Browse all datasets