Even if the two types of interfaces share strong similarities in their most commonplace usage, they are nevertheless accessed through two distinct mediums (from which they get their names (TUI / VUI)). Text is not speech, and speech is not text, and the possibility space of what can be realistically (in a human-friendly manner) provided as input or output on each is quite different. On VUIs, for instance, you can do things like query for a song by humming it [0], which would be very difficult to provide a purely textual alternative for for use in a TUI, whereas text-based interfaces make it easier to do things like looking at large swathes of structured content or copypasting a long piece of text to use as input.

[0]: https://blog.google/products/search/hum-to-search/