It is built so that you edit rather than rewrite. Three things do that work: the voice profile, so it arrives in your register rather than a generic one; the research step, so its claims come from live sources rather than the model’s recollection; and the content check, which names the weakest passages before you have read a word. What lands is a draft with your fingerprints already on it.

Every draft starts with a research step that runs a live search and builds a brief from what it finds, so the article is written against real sources and against what is already ranking for the topic. The brief travels with the draft, so you can see what it was working from.

Google’s stated position is that it rewards helpful content regardless of how it was produced, and penalises content made primarily to game rankings — the same bar a thin human-written post fails. What you get here for that is the research step, the content score and the answer-engine scoring, all of which push a draft towards the first category.

Yes, if that person has written enough for it to read. The voice profile is built from posts you nominate, so pointing it at one author’s work gives you that author’s register — and personas let you publish under that byline. Each blog keeps its own profile, so two blogs on one account can sound nothing alike.