CST: Mobile Voice Crawler

The main articles are:

• Speech Recognition (Technology, Market)
• Speech technology: On the path from recognition to understanding

2024: Mobile Voice Crawler Presentation

The Speech Technology Cent er Group has developed a mobile voice crawler - a solution based on artificial intelligence to automate scheduled maintenance and technical inspections at transport facilities in various industries. The company announced this on July 17, 2024. The AI solution supports work on computers or tablets of specialists, simplifies routine work by voice filling in forms, reports and checklists.

The mobile voice crawler reduces the time for transmitting information about technical inspections by more than five times by optimizing processes at each stage. Earlier, the car crawler manually recorded the required parameters, then transmitted the data to the operator, and the operator entered the received data into the system. With the introduction of an AI solution, it is enough for a specialist to voice the results of the inspection of the car, the voice is converted into text, providing a correctly filled out document of the inspection results without the need to involve third-party specialists.

When developing a mobile voice crawler, we used artificial intelligence, speech recognition and synthesis technologies. So, for example, the high quality of speech recognition technologies made it possible to achieve that the AI solution supports work in difficult acoustic conditions, and high-quality and continuous further training of the neural network helped to achieve excellent recognition even of highly specialized terminology. At the same time, the latest generation of speech synthesis allows, if necessary, to voice the tasks assigned to the employee. The introduction of a mobile voice crawler can significantly speed up typical inspection operations through voice input, ensuring routine control of recorded information, and we welcome its scaling, said Dmitry Dyrmovsky, CEO of the CST Group of Companies.

This solution allows you to increase the share of recorded characteristics by about 80%: if earlier, manually, it was possible to record 20% of the signs of expulsions, then with this decision it will be possible to fix all 100%. In this case, automatic control of permissible parameters when entering information will make it possible to be sure that the data is entered correctly. Subsequently, the data obtained can be used to predictively analyze equipment failures in order to anticipate them, making an even more significant contribution to the safety and economy of the enterprise.