GMAsia
    🇰🇷南韓·AI 新聞·2026年10月22日·來源: 서울경제

    Daglo Launches Korea's First Voice-Matching Tool That Names Speakers

    內容只提供英文版本

    ActionPower has launched "Voice Matching" on its Daglo AI transcription service, introducing Korea's first voiceprint-based tool for general users that identifies and names speakers in meeting minutes. The company reports the feature achieves about 94% accuracy in actual recording environments, building on a single registered voice sample.

    Daglo's new Voice Matching feature aims to streamline meeting documentation by automating speaker identification. Once a user consents and provides an initial name for a voice, the AI learns its characteristics, subsequently labeling speakers automatically in new recordings. This process is designed to reduce the manual effort involved in attributing comments to specific individuals within meeting records.

    The reported 94% accuracy in speaker identification, measured in actual recording environments, indicates a high level of precision in attributing remarks to the correct speaker. This capability allows users to track specific participants' contributions across multiple meetings or generate summaries of individual comments, enhancing the utility of transcribed records for follow-up.

    ActionPower is expanding its enterprise offerings, including a new Team Plan, and has obtained ISO 27001 certification for information security management. The company plans to develop further features, such as automatically organizing assigned tasks and drafting documents based on meeting content, suggesting an intent to integrate AI voice technology more deeply into corporate collaboration workflows.

    分享這篇文章

    延伸閱讀
    原文報道: 서울경제我們不轉載全文, 閱讀原文 →

    相關文章

    6 則