Macgence/indian-agent-to-us-customer-speech-dataset-in-english-for-finance

About This OTS Dataset

This Off-The-Shelf (OTS) dataset offers a comprehensive collection of audio recordings showcasing conversations between US customer representatives and Indian agents within the financial sector. It is meticulously curated to enhance speech recognition and conversational AI models, specifically tailored to the unique dynamics of international financial service interactions.

Metadata Availability: Insights into Participant Details

Each participant is associated with detailed metadata, including:

• Age
• Gender
• Country
• State
• Dialect
• Domain
• Topic
• Call type
• Call outcome

This rich metadata supports effective and informed AI model training and evaluation.

Audio Recording Details

• Audio Duration: 50 hours
• Format: WAV (high-fidelity audio)
• Sample Rate: Variable and customizable
• Bit Depth: Standardized at 16-bit for clarity
• Recording Environments: Captured across various real-world settings
• Equipment Used: Standard call center devices

These specifications ensure compatibility with a wide range of AI development applications in the finance domain.

Insights into Audio Data

This dataset captures 50 hours of natural English audio, focused on financial topics. Key characteristics:

• Features real-world interactions between Indian agents and native US English speakers
• Reflects authentic American accents, dialects, and call center dialogue styles
• Enables accurate training for speech and language models in finance

Dataset Transcription Details

Each audio file is paired with manual verbatim transcriptions in JSON format, including:

• Speaker-wise segmentation
• Timestamps
• Non-speech events labeled

These transcriptions enable efficient training for automatic speech recognition (ASR) and financial domain conversational AI.

License

This dataset is exclusively curated by Macgence and available for commercial use, enabling scalable and robust development of financial AI applications.

Updates and Customization

• Regular updates with newly recorded audio in fresh scenarios
• Customizable options including sample rate adjustments and bespoke transcriptions

👉 Submit your request to discuss your needs, request pricing, or purchase the dataset.

Keywords

keywords: financial speech dataset, English call center dataset, Indian agent US customer dataset, US English speech dataset, conversational AI for finance, ASR dataset finance, financial customer support audio, speech dataset for banking, English speech corpus, finance domain audio dataset, customer service dataset, English transcribed speech dataset, call center voice recordings, English audio for AI training, cross-cultural conversation dataset, real-world finance speech data, machine learning finance dataset, speech recognition in finance, financial dialogue dataset, multilingual speech corpus finance