Does SEO Still Matter Now That People Ask AI?
Yes, because AI answers are assembled from the same web pages search engines index, so the work that makes a page findable also makes it quotable. What changes is the goal: you are no longer trying to rank in a list, you are trying to be the source an answer is built from.
The question is usually asked as though one replaces the other. In practice they share almost all of the same plumbing, and the businesses that abandoned SEO on the assumption that AI made it irrelevant have simply made themselves harder to retrieve.
What does SEO still do that AI has not changed?
It makes your pages retrievable, which is the precondition for everything else. If a page is not crawlable, not indexed, or not linked to from anywhere, it cannot be surfaced by a search engine or used as a source by an assistant.
- Crawlable, indexable pages
- A working sitemap so content can be discovered at all
- Internal links, so pages are not stranded
- Fast, accessible pages that do not fail to load for a bot
- Clear titles and descriptions that state what a page covers
What genuinely changes with AI search?
The unit of success moves from a ranking position to a mention inside an answer, and there is far less room. Being tenth on a results page still got you some traffic. Being the fourth-best source for an AI answer usually gets you nothing.
That compression is why the practical emphasis shifts from keyword coverage toward being the clearest, most corroborated answer on a narrower set of questions.
Does traffic go down even when visibility goes up?
It can, and this is the part businesses find hardest. If an assistant answers the question completely, the user may never click, so you can be more visible and less trafficked at the same time.
The practical response is to stop treating sessions as the only measure of success, and to watch brand searches, direct visits and inbound enquiries that arrive already informed. More on that in measuring whether AI search sends customers.
Should you write differently now?
Write so that any single section can be lifted and still make sense. Lead with the answer, keep claims specific and attributable, and structure around genuine questions rather than clever headings.
This tends to make content better for humans as well, which is the rare case where optimising for machines and for readers point the same direction.
What about keywords?
They still matter as evidence of what people want to know, but they are a worse target than the question behind them. Optimising a page for a phrase produces content that repeats the phrase. Optimising for the question produces content that answers it, and answers are what get used.
So where should effort go?
Keep the technical foundation, then spend the marginal hour on being the best answer to a smaller number of questions your buyers actually ask. For how that translates into being named, see how businesses get recommended by AI.
Engage Ads builds paid advertising, AI search visibility and lead follow-up as one system for high-ticket service businesses. If you want to see where yours is leaking, start with a no-cost growth audit.
