One of the issues that I discovered in macOS Sequoia v15.6.1 is that Apple's Shortcut action to transcribe dictated voice audio files (e.g. .m4a) to text fails before even a paragraph of text has been transcribed.
Whether on Intel, or my M4 Mac Mini Pro, I can convert three paragraphs of text to a M4A audio file and when transcribed, the last two paragraphs and last sentence of the first paragraph are absent in the resulting text. Upon listening to the audio file, all content is present. The English voice performing the Text to Voice activity is clearly understood in the audio content. This suggests that you may need a competent third-party transcirption tool to completely process the audio file.
In my own experience with dictating directly into Pages with a decent microphone, this will create more lost productivity fighting with the dictation process than if you just resigned yourself to directly type the content into Pages. Dictation does not tolerate pauses in your speech while you think about what to say next. Your mileage may vary though… 😉