Social chat might be taking a different direction. A microphone icon can conceal two very different experiences. In one, a visitor speaks a question, watches it become text, and waits for a written answer. In the other, the visitor has a spoken exchange, adding details or interrupting a reply as the conversation develops. Both may be called voice chat, but the distinction affects everything from interface design to software architecture.
Voice input is primarily an alternative to typing. Speech recognition converts audio into words, which are passed to a chatbot much like a keyboard submission. The assistant can respond on screen, and a separate text-to-speech component may read that response aloud. This approach remains useful: it gives developers clear transcripts and considerable control over each processing stage.
Adding spoken output does create a voice conversation, so the terminology needs care.…