Bitext - Travel Tagged Training Dataset for LLM-based Virtual Assistants

  Overview

This hybrid synthetic dataset is designed to be used to fine-tune Large Language Models such as GPT, Mistral and OpenELM, and has been generated using our NLP/NLG technology and our automated Data Labeling (DAL) tools. The goal is to demonstrate how Verticalization/Domain Adaptation for the [Travel] sector can be easily achieved using our two-step approach to LLM Fine-Tuning. An overview of… See the full description on the dataset page: https://huggingface.co/datasets/bitext/Bitext-travel-llm-chatbot-training-dataset.

Bitext - Customer Service Tagged Training Dataset for LLM-based Virtual Assistants

Overview

This dataset can be used to train Large Language Models such as GPT, Llama2 and Falcon, both for Fine Tuning and Domain Adaptation.

The dataset has the following specs:

• Use Case: Intent Detection
• Vertical: Customer Service
• 27 intents assigned to 10 categories
• 26872 question/answer pairs, around 1000 per intent
• 30 entity/slot types
• 12 different types of language generation tags

The categories and intents have been selected from Bitext's collection of 20 vertical-specific datasets, covering the intents that are common across all 20 verticals. The verticals are:

• Automotive, Retail Banking, Education, Events & Ticketing, Field Services, Healthcare, Hospitality, Insurance, Legal Services, Manufacturing, Media Streaming, Mortgages & Loans, Moving & Storage, Real Estate/Construction, Restaurant & Bar Chains, Retail/E-commerce, Telecommunications, Travel, Utilities, Wealth Management

For a full list of verticals and its intents see https://www.bitext.com/chatbot-verticals/.

The question/answer pairs have been generated using a hybrid methodology that uses natural texts as source text, NLP technology to extract seeds from these texts, and NLG technology to expand the seed texts. All steps in the process are curated by computational linguists.

Dataset Token Count

The dataset contains an extensive amount of text data across its 'instruction' and 'response' columns. After processing and tokenizing the dataset, we've identified a total of 3.57 million tokens. This rich set of tokens is essential for training advanced LLMs for AI Conversational, AI Generative, and Question and Answering (Q&A) models.

Fields of the Dataset

Each entry in the dataset contains the following fields:

• flags: tags (explained below in the Language Generation Tags section)
• instruction: a user request from the Customer Service domain
• category: the high-level semantic category for the intent
• intent: the intent corresponding to the user instruction
• response: an example expected response from the virtual assistant

Categories and Intents

The categories and intents covered by the dataset are:

• ACCOUNT: create_account, delete_account, edit_account, recover_password, registration_problems, switch_account
• CANCELLATION_FEE: check_cancellation_fee
• CONTACT: contact_customer_service, contact_human_agent
• DELIVERY: delivery_options, delivery_period
• FEEDBACK: complaint, review
• INVOICE: check_invoice, get_invoice
• ORDER: cancel_order, change_order, place_order, track_order
• PAYMENT: check_payment_methods, payment_issue
• REFUND: check_refund_policy, get_refund, track_refund
• SHIPPING_ADDRESS: change_shipping_address, set_up_shipping_address
• SUBSCRIPTION: newsletter_subscription

Entities

The entities covered by the dataset are:

• {{Order Number}}, typically present in:
• Intents: cancel_order, change_order, change_shipping_address, check_invoice, check_refund_policy, complaint, delivery_options, delivery_period, get_invoice, get_refund, place_order, track_order, track_refund
• {{Invoice Number}}, typically present in:
  • Intents: check_invoice, get_invoice
• {{Online Order Interaction}}, typically present in:
  • Intents: cancel_order, change_order, check_refund_policy, delivery_period, get_refund, review, track_order, track_refund
• {{Online Payment Interaction}}, typically present in:
  • Intents: cancel_order, check_payment_methods
• {{Online Navigation Step}}, typically present in:
  • Intents: complaint, delivery_options
• {{Online Customer Support Channel}}, typically present in:
  • Intents: check_refund_policy, complaint, contact_human_agent, delete_account, delivery_options, edit_account, get_refund, payment_issue, registration_problems, switch_account
• {{Profile}}, typically present in:
  • Intent: switch_account
• {{Profile Type}}, typically present in:
  • Intent: switch_account
• {{Settings}}, typically present in:
  • Intents: cancel_order, change_order, change_shipping_address, check_cancellation_fee, check_invoice, check_payment_methods, contact_human_agent, delete_account, delivery_options, edit_account, get_invoice, newsletter_subscription, payment_issue, place_order, recover_password, registration_problems, set_up_shipping_address, switch_account, track_order, track_refund
• {{Online Company Portal Info}}, typically present in:
  • Intents: cancel_order, edit_account
• {{Date}}, typically present in:
  • Intents: check_invoice, check_refund_policy, get_refund, track_order, track_refund
• {{Date Range}}, typically present in:
  • Intents: check_cancellation_fee, check_invoice, get_invoice
• {{Shipping Cut-off Time}}, typically present in:
  • Intent: delivery_options
• {{Delivery City}}, typically present in:
  • Inten...

The University Chatbot Dataset contains 38 intents covering general university-related inquiries, designed to train, fine-tune, and evaluate conversational AI models in the education sector.

Bitext - Retail Banking Tagged Training Dataset for LLM-based Virtual Assistants

  Overview

This hybrid synthetic dataset is designed to be used to fine-tune Large Language Models such as GPT, Mistral and OpenELM, and has been generated using our NLP/NLG technology and our automated Data Labeling (DAL) tools. The goal is to demonstrate how Verticalization/Domain Adaptation for the [Retail Banking] sector can be easily achieved using our two-step approach to LLM Fine-Tuning.… See the full description on the dataset page: https://huggingface.co/datasets/bitext/Bitext-retail-banking-llm-chatbot-training-dataset.

This JSON file contains a collection of conversational AI intents designed to motivate and interact with users. The intents cover various topics, including greetings, weather inquiries, hobbies, music, movies, farewells, informal and formal questions, math operations and formulas, prime numbers, geometry concepts, math puzzles, and even a Shakespearean poem.

The additional intents related to consolidating people and motivating them have been included to provide users with uplifting and encouraging responses. These intents aim to offer support during challenging times, foster teamwork, and provide words of motivation and inspiration to users seeking guidance and encouragement.

The JSON structure is organized into individual intent objects, each containing a tag to identify the intent, a set of patterns representing user inputs, and corresponding responses provided by the AI model. This dataset can be used to train a conversational AI system to engage in positive interactions with users and offer motivational messages.

Bitext Sample Pre-built Customer Support Dataset for English

Overview

This dataset contains example utterances and their corresponding intents from the Customer Support domain. The data can be used to train intent recognition models Natural Language Understanding (NLU) platforms.

The dataset covers the "Customer Support" domain and includes 27 intents grouped in 11 categories. These intents have been selected from Bitext's collection of 20 domain-specific datasets (banking, retail, utilities...), keeping the intents that are common across domains. See below for a full list of categories and intents.

Utterances

The dataset contains over 20,000 utterances, with a varying number of utterances per intent. These utterances have been extracted from a larger dataset of 288,000 utterances (approx. 10,000 per intent), including language register variations such as politeness, colloquial, swearing, indirect style... To select the utterances, we use stratified sampling to generate a dataset with a general user language register profile.

The dataset also reflects commonly ocurring linguistic phenomena of real-life chatbots, such as: - spelling mistakes - run-on words - missing punctuation

Contents

Each entry in the dataset contains an example utterance from the Customer Support domain, along with its corresponding intent, category and additional linguistic information. Each line contains the following four fields: - flags: the applicable linguistic flags - utterance: an example user utterance - category: the high-level intent category - intent: the intent corresponding to the user utterance

Linguistic flags

The dataset contains annotations for linguistic phenomena, which can be used to adapt bot training to different user language profiles. These flags are: B - Basic syntactic structure S - Syntactic structure L - Lexical variation (synonyms) M - Morphological variation (plurals, tenses…) I - Interrogative structure C - Complex/Coordinated syntactic structure P - Politeness variation Q - Colloquial variation W - Offensive language E - Expanded abbreviations (I'm -> I am, I'd -> I would…) D - Indirect speech (ask an agent to…) Z - Noise (spelling, punctuation…)

These phenomena make the training dataset more effective and make bots more accurate and robust.

Categories and Intents

The intent categories covered by the dataset are: ACCOUNT CANCELLATION_FEE CONTACT DELIVERY FEEDBACK INVOICES NEWSLETTER ORDER PAYMENT REFUNDS SHIPPING

The intents covered by the dataset are: cancel_order complaint contact_customer_service contact_human_agent create_account change_order change_shipping_address check_cancellation_fee check_invoices check_payment_methods check_refund_policy delete_account delivery_options delivery_period edit_account get_invoice get_refund newsletter_subscription payment_issue place_order recover_password registration_problems review set_up_shipping_address switch_account track_order track_refund

(c) Bitext Innovations, 2020

Overview This dataset is designed to train and fine-tune chatbot models by mapping user queries (patterns) to predefined intents (tags) and generating contextually accurate responses. Each tag represents a unique conversational intent or topic (e.g., "climate_change," "crypto_regulation," "quantum_computing"), accompanied by multiple paraphrased user prompts (patterns) and a detailed, informative response. Ideal for building intent classification systems, dialogue management, or generative AI models.

{ "intents": [ { "tag": "tag_name", "patterns": ["user query 1", "user query 2", ...], "responses": ["detailed answer"] }, ... ] }

Possible Uses Intent Classification: Train models to categorize user inputs into predefined tags.

Response Generation: Fine-tune generative models (GPT, BERT) to produce context-aware answers.

Educational Chatbots: Power QA systems for topics like science, history, or technology.

Customer Support: Automate responses for FAQs or policy explanations.

Compatibility Frameworks: TensorFlow, PyTorch, spaCy, Rasa, Hugging Face Transformers.

Use Cases: Virtual assistants, customer service bots, trivia apps, educational tools.

We’ve developed another annotated dataset designed specifically for conversational AI and companion AI model training.

Watch: How To Use The Dataset

What you have here on Kaggle is our free sample - Think Salon Kitty meets AI

The 'Time Waster Identification & Retreat Model Dataset', enables AI handler agents to detect when users are likely to churn—saving valuable tokens and preventing wasted compute cycles in conversational models.

This batch has 167 entries annotated for sentiment, intent, user risk flagging (via behavioural tracking), user Recovery Potential per statement; among others. This dataset is designed to be a niche micro dataset for a specific use case: Time Waster Identification and Retreat.

👉 Buy the updated version: https://lifebricksglobal.gumroad.com/l/Time-WasterDetection-Dataset

This dataset is perfect for:

• Fine-tuning LLM routing logic
• Building intelligent AI agents for customer engagement
• Companion AI training + moderation modelling
• This is part of a broader series of human-agent interaction datasets we are releasing under our independent data licensing program.

It is designed for AI researchers and developers building:

• Conversational AI agents
• Companion AI models
• Human-agent interaction simulators
• LLM routing optimization models

Use case:

• Conversational AI
• Companion AI
• Defence & Aerospace
• Customer Support AI
• Gaming / Virtual Worlds
• LLM Safety Research
• AI Orchestration Platforms

This batch has 167 entries annotated for sentiment, intent, user risk flagging (via behavioural tracking), user Recovery Potential per statement; among others. This dataset is designed to be a niche micro dataset for a specific use case: Time Waster Identification and Retreat.

👉 Good for teams working on conversational AI, companion AI, fraud detectors and those integrating routing logic for voice/chat agents

👉 Buy the updated version: https://lifebricksglobal.gumroad.com/l/Time-WasterDetection-Dataset

Contact us on LinkedIn: Life Bricks Global.

License:

This dataset is provided under a custom license. By using the dataset, you agree to the following terms:

Usage: You are allowed to use the dataset for non-commercial purposes, including research, development, and machine learning model training.

Modification: You may modify the dataset for your own use.

Redistribution: Redistribution of the dataset in its original or modified form is not allowed without permission.

Attribution: Proper attribution must be given when using or referencing this dataset.

No Warranty: The dataset is provided "as-is" without any warranties, express or implied, regarding its accuracy, completeness, or fitness for a particular purpose.

Bitext - Restaurants Tagged Training Dataset for LLM-based Virtual Assistants

  Overview

This hybrid synthetic dataset is designed to be used to fine-tune Large Language Models such as GPT, Mistral and OpenELM, and has been generated using our NLP/NLG technology and our automated Data Labeling (DAL) tools. The goal is to demonstrate how Verticalization/Domain Adaptation for the [restaurants] sector can be easily achieved using our two-step approach to LLM Fine-Tuning. An… See the full description on the dataset page: https://huggingface.co/datasets/bitext/Bitext-restaurants-llm-chatbot-training-dataset.

Dataset

This dataset was created by Abhishek Srivastava

Contents

📌 Overview This dataset is designed to support research in AI-driven language learning, specifically for chatbot-based English tutoring. It includes intent classification for chatbot interactions and grammatical error correction to assist users in improving their English proficiency.

📊 Dataset Structure The dataset consists of 200 rows with the following columns:

Sentence → User queries for intent classification (e.g., "Can you check my grammar?") Intent → Categorized chatbot responses (e.g., Grammar_Check, Vocabulary_Assistance) Incorrect_Sentence → Common grammatical errors in English writing Corrected_Sentence → AI-corrected versions of the incorrect sentences

[EN] French training dataset for chatbots dealing with usual requests on bank cards.

• Description: This dataset represents examples of common customer requests relating to bank cards management. It can be used as a training set for a small chatbot intended to process these usual requests.
• Content: The questions are asked in French. The dataset is divided into 10 intents of 50 questions each, for a total of 500 questions.
• Intents scope: Intents are constructed in such a way that all questions arising from the same intention have the same response or action. The scope covered concerns: loss or theft of cards; the swallowed card; the card order; consultation of the bank balance; insurance provided by a card; card unlocking; virtual card management; management of bank overdraft; management of payment limits; management of contactless mode.
• Origin: Intents scope is inspired by a chatbot currently in production, and the wording of the questions are inspired by the usual customers requests.

[FR] Jeu d'entraînement en français d'assistants conversationnels traitant des demandes courantes sur les cartes bancaires.

• Description : Cet ensemble de données représente des exemples de demandes usuelles des clients concernant la gestion des cartes bancaires. Il peut être utilisé comme jeu d'entraînement pour un assistant conversationnel destiné à traiter ces demandes courantes.
• Contenu : Les questions sont formulées en français. L'ensemble de données est divisé en 10 intentions de 50 questions chacune, pour un total de 500 questions.
• Périmètre des intentions : Les intentions sont construites de telle manière que toutes les questions issues d'une même intention ont la même réponse ou action. Le périmètre couvert concerne : la perte ou le vol de cartes ; la carte avalée ; la commande des cartes ; la consultation du solde bancaire ; l'assurance fournie par une carte ; le déverrouillage de la carte ; la gestion de cartes virtuelles ; la gestion du découvert bancaire ; la gestion des plafonds de paiement ; la gestion du mode sans contact.
• Origine : Le périmètre des intentions est inspiré par un chatbot actuellement en production, et la formulation des questions est inspirée de demandes courantes de clients.

Context

I was working with RNN models in Tensorflow and was searching about conversation bots. Then a idea struck me as to create a bot myself. I looked for chat data but was not able to find something useful. Then I came across Meena chatbot and Mitsoku chatbot data and so compiled them with some data from human chats corpus.

Content

The data corpus contain chat labelled chat data with Human 1 and Human 2 in ask-reponse manner. Each odd row with Human 1 label is the initiator of the chat and each even row with Human 2 label is the response. Data after Human x: is the chat data which can be preprocessed to remove the label part.

Acknowledgements

We wouldn't be here without the help of others. If you owe any attributions or thanks, include them here along with any citations of past research.

Inspiration

I would love others to explore this data and frame ideas related to the creation of a chatbot system.

Ethical Clearance Reference: 216153972/2024/2This dataset presents the thematic analysis of the qualitative methodology to identify the factors affecting the diffusion of AI technologies for the purposes of customer service training improvement. The data were collected using semi-structured interviews with participants from a purposive sample. This was then transcribed and anonymised to aid the thematic analysis for the purpose of the study. Key themes were highlighted in the data and revealed that AI-enabled chatbots improved information accessibility, offered personalised learning opportunities, facilitated self-paced and adaptive learning, provided consistency in responses, enhanced operational efficiency among customer service agents, and contributed to teamwork and engagement. Findings suggest that the AI tool is most effective when used in conjunction with human facilitation. The dataset highlights both the benefits and limitations of using AI-enabled tools in training environments. Challenges included the limited content depth of the responses and technical infrastructure constraints that questioned organisational readiness and strategic direction. The dataset was used to derive a framework to support the effective implementation and integration of AI tools to enhance customer service training.

Snapshot img

Chatbot Market Size 2025-2029

The chatbot market size is forecast to increase by USD 9.63 billion, at a CAGR of 42.9% between 2024 and 2029. Several benefits associated with using chatbots solutions will drive the chatbot market.

Major Market Trends & Insights

APAC dominated the market and accounted for a 37% growth during the forecast period.
By End-user - Retail segment was valued at USD 210.60 billion in 2023
By Product - Solutions segment accounted for the largest market revenue share in 2023

Market Size & Forecast

Market Opportunities: USD 1.00 billion
Market Future Opportunities: USD 9.63 billion 
CAGR : 42.9%
APAC: Largest market in 2023

Market Summary

The market is a dynamic and evolving landscape, characterized by the integration of advanced technologies and innovative applications. Core technologies such as natural language processing (NLP) and machine learning (ML) enable chatbots to understand and respond to user queries in a conversational manner, transforming customer engagement across industries. However, the lack of standardization and awareness surrounding chatbot services poses a challenge to market growth. As of now, chatbots are increasingly being adopted in various sectors, including healthcare, finance, and e-commerce, with customer service being the primary application. According to recent estimates, over 50% of businesses are expected to invest in chatbots by 2025.
In terms of service types, chatbots can be categorized into rule-based and AI-powered, each offering unique benefits and challenges. Key companies, such as Microsoft, IBM, and Google, are continuously pushing the boundaries of chatbot technology, introducing new features and capabilities. Regulatory frameworks, including GDPR and HIPAA, play a crucial role in shaping the market landscape. Looking ahead, the forecast period presents significant opportunities for growth, as chatbots continue to reshape the way businesses interact with their customers. Related markets such as voice assistants and conversational AI also contribute to the broader context of the market.
Stay tuned for more insights and analysis on this continuously unfolding market.

What will be the Size of the Chatbot Market during the forecast period?

Get Key Insights on Market Forecast (PDF) Request Free Sample

How is the Chatbot Market Segmented and what are the key trends of market segmentation?

The chatbot industry research report provides comprehensive data (region-wise segment analysis), with forecasts and estimates in 'USD million' for the period 2025-2029, as well as historical data from 2019-2023 for the following segments.

End-user

  Retail
  BFSI
  Government
  Travel and hospitality
  Others


Product

  Solutions
  Services


Deployment

  Cloud-Based
  On-Premise
  Hybrid


Application

  Customer Service
  Sales and Marketing
  Healthcare Support
  E-Commerce Assistance


Geography

  North America

    US
    Canada


  Europe

    France
    Germany
    Italy
    UK


  Middle East and Africa

    Egypt
    KSA
    Oman
    UAE


  APAC

    China
    India
    Japan


  South America

    Argentina
    Brazil


  Rest of World (ROW)

By End-user Insights

The retail segment is estimated to witness significant growth during the forecast period.

The market is experiencing significant growth, with adoption in various sectors escalating at a remarkable pace. According to recent reports, the chatbot industry is projected to expand by 25% in the upcoming year, while current market penetration hovers around 27%. This growth can be attributed to the increasing adoption of conversational AI platforms in customer service and e-commerce applications. Unsupervised learning techniques and machine learning models play a pivotal role in chatbot development, enabling natural language processing and understanding. Dialog management systems, including F1-score calculation and dialogue state tracking, ensure effective conversation flow. Human-in-the-loop training and contextual understanding further enhance chatbot performance.

Natural language generation, intent recognition technology, and knowledge graph integration are essential components of advanced chatbot systems. Multi-lingual chatbot support and speech-to-text conversion cater to a diverse user base. Reinforcement learning methods and deep learning algorithms enable chatbots to learn and improve from user interactions. Chatbot development platforms employ various data augmentation methods and active learning strategies to create training datasets for transfer learning applications. Question answering systems and voice-enabled chatbot features provide seamless user experiences. Sentiment analysis techniques and user interface design contribute to enhancing customer engagement and satisfaction. Conversational flow design and response generation models ensure e

Victorano/Bitext-customer-support-llm-chatbot-training-dataset-4k-seed42 dataset hosted on Hugging Face and contributed by the HF Datasets community

Get a high-quality chatbot dataset for AI/ML models in BFSI Sector. Train with diverse conversational data for accurate, efficient machine learning applications

Given the importance of conversation practice in language learning, chatbots, especially ChatGPT, have attracted considerable attention for their ability to converse with learners using natural language. This review contributes to the literature by examining the currently unclear overall effect of using chatbots on language learning performance and comprehensively identifying important study characteristics that affect the overall effectiveness. We meta-analyzed 70 effect sizes from 28 studies, using robust variance estimation. The effects were assessed based on 18 study characteristics about learners, chatbots, learning objectives, context, communication/interaction, and methodological and pedagogical designs. Results indicated that using chatbots produced a positive overall effect on language learning performance (g = 0.486), compared to non-chatbot conditions. Moreover, four characteristics (i.e., educational level, language level, interface design, and interaction capability) affected the overall effectiveness. In an in-depth discussion on how the 18 characteristics are related to the effectiveness, future implications for practice and research are presented.

This dataset provides detailed, synthetic healthcare chatbot conversations with annotated intent labels, message sequencing, and extracted entities. Designed for training and evaluating conversational AI, it supports intent classification, dialogue modeling, and entity recognition in healthcare virtual assistants. The dataset enables robust analysis of user-bot interactions for improved patient engagement and automation.

Get a high-quality chatbot dataset for AI/ML models in Hospitality Sector. Ideal for NLP training, improving chatbot responses, and enhancing conversational AI.