If you're a not Project Fi user, or one who avidly uses Hangouts, you may still be using Google Voice. Even if you use Hangouts quite often, Google Voice may still be an installed app on your device, and that's probably because it still holds some use to you. One of the more popular features of Google Voice has always been its ability to transcript voicemails into text, the only downside is that those transcripts weren't always the most easy to understand, and in some cases they were just downright terrible, feeding you a message with half legible text while the other half of the message read like someone who was quite confused wrote the message.
Google has improved Voice much over the years, but as of now it seems Google has added one very requested improvement that users have been more or less begging for, and that's better transcription of the voicemails which come through to Google Voice numbers as Google highlighted in an earlier blog post. Users could always get around this of course by just listening to the voicemail instead of reading the transcription, but the benefit of being able to read it was such that it could be viewed at times when perhaps a voicemail could not be listened to due to possible interruption of what may have been happening at any given time.
After so long, Google begun asking users to share some of their voicemail transcriptions to help them research the issue and make improvements, and they have finally made some fine-tuned adjustments and have brought the transcription errors down by a staggering 49 percent. That's less than almost half of the errors that were there prior to these updates, which is likely going to make plenty of users very happy. Those who still use Google Voice of course. Good news for Project Fi users too though, is that these voicemail transcription improvements apply to the Project Fi as well, so you should be receiving many less errors when people leave voicemails for you down the line. Google mentions that much of this improvement should be attributed to the users who shared transcriptions, and with continued support they hope to make the whole process even better. Perhaps one day, errors will be almost nonexistent.