Splitting AI search and traditional search into two teams produces two plans that fight each other. They draw on the same inputs: the same pages, the same authority, the same crawl.
What they share
Assistants mostly answer by searching. The pages they pull are, in large part, pages that rank. Being findable in classic search is still the entry ticket, and the work that makes a page rank is usually the work that makes it quotable.
Where they genuinely differ
The divergence is real and worth naming. Assistants rewrite the question into their own searches, so the query you optimised for is often not the query that gets run. They read passages rather than pages. And they pick a handful of sources, so the long tail of positions four to ten matters far less than it used to.
Our own testing keeps finding the same thing: ranking gets you considered, and something else gets you chosen.
Table stakes
- One owner for organic, covering both.
- One content calendar, not two.
- Shared reporting where classic rank and AI recommendation sit on the same page.
- A test habit: change one thing, keep a control, look again in a month.
Run them apart and you will spend twice to learn the same lessons.