Industry Solutions
Financial Services
Financial Services
Streamling your business process with enterprise-level security guaranteed.
Our top priority lays on compliance and security.
Revitalizing financial industry by providing one-stop AI data solutions for customer service, virtual counter, virtual assistant, targeting marketing, and other AI applications.
Scenarios
Customer Service
Dial In/Out, Collect Accounts Receivable
Smart Meeting
Realtime Captioning,Translation,Meeting Minutes Generation
Automated Invoice Processing
Identity Certification, Warranty OCR, Medical Record OCR
Virtual Human
Smart Shopping Guidance, Marketing
Challenge
Imprecise voice recognition of customer service scenario
Unable to correctly understand commands and queries
Impersonal and unnatural communication
Different format between invoices, warranties, and medical records
Annotator® AI-Assisted Annotation Platform
Audio Annotation
Text Annotation
Image Annotation
- Customer Service - Customer service annotation
- Virtual Human - Command and query annotation
- Smart Meeting - Meeting scenarios voice annotation
- Virtual Human - Rhythm, Text segmentation, part-of-speech, and phoneme annotation
- Customer Service - User queries relevance annotation
- Virtual Human - User interaction content annotation
- Automated Invoice Processing - Invoice OCR annotation
MD Dataset Portfolio
Speech Recognition
→
Text-to-Speech
→
Natural Language Understanding
→
OCR
→
Contact us for data collection and annotation service
Related Datasets
MDT-AF063 Mandarin Chinese Scripted Speech Corpus
Play Audio
MDT-NG001 Chinese POI Text Corpus
MDT-AF069 English Duplex Conversation Training Dataset
Magic Data has introduced the "Multi-stream Spontaneous Conversation Training Datasets_English".This dataset comprises 5,000 hours of multi-accent English conversational data, encompassing a wide range of vocal scenarios. Our dataset allows AI models to better understand contextual changes, tonal variations, and emotional shifts in conversations, thereby producing responses that are more natural and accurate.
Play Audio
MDT-NB002 Chinese Named Entity Extraction (NEE) Corpus
MDT-RI002 Cantonese Spoken Speech Dataset
This dataset is built to train AI models that better understand spoken Cantonese, improving natural interaction in speech recognition. It covers diverse real-life dialogues with high transcription accuracy. Annotations are optimized for Cantonese-specific features like nine tones, lazy sounds, and slang, ensuring accurate audio-text alignment. Natural sentence structures and punctuation help models grasp pause patterns and sentence-final particles, boosting dialect recognition performance.
MDT-AF012 Singaporean English Scripted Speech Corpus—Daily Use Sentence
Play Audio