Kakao Enterprise Unveils Customer-Customized AI Voice Recognition Service 'Custom STT'

[Asia Economy Reporter Kang Nahum] Kakao Enterprise unveiled its enterprise AI speech recognition conversion model, 'Custom STT (Custom Speech-to-Text),' on the 25th.

STT is a function that uses AI to convert speech into text and is utilized for various purposes such as AI speakers, voice search, and navigation.

The newly released Custom STT allows companies to easily and quickly build models tailored to their customers and industry characteristics, improving the accuracy of speech recognition conversion. When clients add words and patterns themselves, the most suitable AI model is automatically created in a short time. It can also handle proper nouns and long, complex words.

A Kakao Enterprise official stated that internal company experiments showed a more than 30% improvement in speech recognition error rates compared to existing models.

As a cloud-based 'AI as a Service (AIaaS),' it also has the advantage of reducing the time and cost burden of service management for companies. To modify or develop existing STT services, clients must organize their requests and deliver them to developers, who then have to redevelop the model.

The Custom STT model allows clients to directly access the cloud and input new content such as product names, new features, and requests in real time to customize, significantly reducing time and costs.

Kakao Enterprise plans to apply Custom STT to inquiry response and consultation tasks at the 'Kakao i Connect Center (AICC),' scheduled to open in the first half of this year. Choi Dongjin, head of Kakao Enterprise AI Lab, said, "In the future, we plan to sequentially release our proprietary technologies in fields such as optical character recognition (OCR) and text-to-speech (TTS) as customizable AI as a Service (AIaaS) models, not only speech recognition."

This content was produced with the assistance of AI translation services.