We learned through testing that speech is a unique form of communication presenting interesting challenges. Some people became flustered when speaking into a mic and some felt too self-conscious to want to do it all. It turned out that the quality of transcription was good enough to get by, but the issue was user behavior.
On the whole, we found it was a unique individual that enjoyed communicating in this way. It required a level of intimacy in your relationship, so talking to people you're very close to in real life, and people you regularly communicate with. In those situations, Howdy offered a much faster and more personal way to stay in touch often, especially while on the go such as in the car. However, we found that it was still slightly too niche and the app wasn't fully featured enough to drive users from their current communication channel.
That said, very recently we've seen improvements to voice-to-text from Apple
as well as Telegram
who added transcription as a premium feature to voice messages. So I definitely think voice-to-text is still an area ripe for innovation and improvement.