DX.YI - Intelligent Al voice platform

Product Overview

The intelligent AI voice platform integrates large model technology and cutting-edge technologies such as ASR(speech recognition), TTS(speech synthesis) and NLP(natural language processing), adopts voice print recognition and visual process editing functions, and flexibly connects through standardized open interfaces to Provide highly anthropomorphic intelligent voice outbound calling, intelligent customer service, and intention recognition services.The application of this comprehensive technology platform significantly reduces the operator's labor costs, optimizes the complaint handling and marketing outbound call workflow, improves the efficiency of analysis and judgment, and brings a qualitative leap for the operator's business operation.

Technical framework

Based on the outbound big data of operators, the platform realizes the automatic creation of outbound task data, and uses A1 big model to complete automatic classification. Through the pre-set time, order and call technique, the system makes outgoing calls in order, automatically records and converts the speech into text form. At the same time, the platform can automatically recognize and classify speech based on speech design and speech features in real time, and establish corresponding groups and result libraries. This enables the platform to automatically complete more than 80% of daily operational work such as marketing recommendations and complaint handling while conducting subsequent human intervention.

Product function

Automatic identificationof speech analysis

Develop different scripts based on different scenarios, train intelligent AI robots according to their application scenarios, and ultimately achieve automatic identification and tagging of specific fields with corresponding labels.

AI intelligent outbound calling

It can automatically generate various types of robot synthesized speech for well-made scripts, or manually record them in batches. Upload the voice, configure the corresponding scripts, and form a complete voice robot. This robot can be repeatedly set up multiple times.

Complaint handling

In response to operator complaints, the robot responds intelligently, records relevant complaint information, recognizes complaint classification, and marks malicious complaints and other complaint-related functions.

AI large model-intention recognition

Based on the information from human-generated outbound calling voice corpuses, a model is trained to accomplish intent recognition and classification processing through advanced algorithmic models.

Voiceprint recognition

Audio preprocessing, audio noise reduction and speech enhancement, audio effective sound extraction, live body detection, audio quality detection, voiceprint feature extraction, voiceprint feature comparison, voiceprint library management and encryption and other functions.

Voice call to text conversion

Convert voice call content into text content and store it.