The speech recognition service is now slightly more capable, and as you continue with the meeting, the initial recognition will be more accurate. When you replay the meeting, you should now see more accurate subtitles!
The previous basic model's speech recognition accuracy was not very high. As a result, we often had to edit the text of voice notes before sharing the contents of meetings. To alleviate these difficulties, we have applied a higher accuracy speech recognition algorithm. This algorithm will be gradually implemented after voice notes are first converted.
Therefore, please proceed with the meeting as usual even if the voice notes are recorded differently at the beginning of the meeting! When reviewing the meeting later, it will be recorded more accurately.