People keep asking me whether AI Overviews killed SEO. They did not, but they did something messier and more interesting: they took one job and split it into two. The old job of ranking is still there, and on top of it you now have to get your name into the answer that sits above the rankings, so you are really doing both at once.
This is the part that trips people up, so let me be plain about it. The old playbook still works, it just stopped being enough on its own. Below is what stayed the same, what came apart, and the formatting that actually earns the citation. This is a supporting piece for my SEO Trends 2026 report, so start there if you want the full year in context.
What stayed the same: you still have to rank
There is something the people selling “GEO courses” tend to skip over. AI Overviews pull from pages Google already trusts. The model is not crawling the open web fresh and forming its own opinion, it is synthesizing from results that already rank, so the pool it cites from is mostly the pool that already ranks.
That makes classic SEO and the AI citation close to the same job right now rather than two competing strategies. The technical hygiene, the topical depth, the backlinks, the first-hand experience, the trust signals, all of it still does the work it always did. A page that cannot rank almost never gets cited, which makes getting to page one the price of admission to the answer box.
So when someone tells you SEO is dead and GEO is the new thing, they are half right and half selling. The fundamentals have not moved. What moved is where the finish line sits.
What’s new: ranking and visibility came apart
For twenty years ranking and visibility amounted to the same thing: rank well and you got seen, and being seen got you the click. AI Overviews broke that chain.
Now you can rank in position three and be completely invisible, because the AI Overview answered the question and nobody scrolled down to you. You can also sit below the top spot and still get named inside the answer, which puts your brand in front of someone before a single blue link loads. Ranking and being cited have become two separate outcomes that no longer travel together.
The numbers make this concrete. AI Overviews appear on roughly half of Google searches1, and when one shows up the click rate on an organic result drops about 61%2. About 68% of Google searches now end with zero clicks3. On the surface that all reads pretty grim.
The part the surface read misses is where the remaining clicks go. Brands cited inside the AI Overview earn around 120% more clicks per impression than uncited brands on the same query2. Same query, same page of results, and the only thing separating them is whether the AI named one of them. The clicks that still happen go to whoever got quoted.
If there is one line to take from this, it is that the click now follows the citation more than the ranking. Ranking still gets you considered, and getting named in the synthesized answer is the part that actually pulls the click.
I wrote up what this does to Google’s own front door in how Google Search itself is trending, because AI Mode going default changes how the question even gets asked.
How to earn the citation
This is the part you can act on this afternoon. The engine wants clean, liftable sentences from pages it already trusts. Give it one and it quotes you, and bury your answer under three paragraphs of throat-clearing and it quotes whoever was clearer instead.
Lead with a quotable answer
Put a tight, direct answer to the page’s core question in the first 40 to 60 words. Not a hook or a setup, just the answer stated cleanly, in a sentence a model can lift whole. If a reader could copy that one sentence and have their question answered, an engine can too. Everything else on the page can stay as long and rich as you like. The top just has to be extractable.
Write question-shaped headings
Use H2s that mirror real People Also Ask phrasing. “How do I get cited in an AI Overview” beats “Citation strategy.” The engine matches the question to your heading, then reads the answer underneath it. Make that match obvious and you make yourself easy to cite.
Add comparison tables
Models love structured data they can lift without rewriting. A clean comparison table is some of the most citable content you can publish.
| Job | Classic SEO | GEO |
|---|---|---|
| Goal | Rank on the page | Get named in the answer |
| Wins the | Position | Citation |
| Format that helps | Depth, links, trust | Quotable answer up top |
| Measured by | Rank tracker | Visibility in AI answers |
Use schema
Mark up your pages with Article, FAQ, HowTo, and Breadcrumb schema. Structured data tells the engine what each chunk of your page is, which makes extraction cleaner and citation more likely. It is not magic, but it is cheap and it removes friction.
Keep cornerstone content fresh
Engines weigh recency. A strong page from 2024 with no updates loses ground to a 2026 page on the same topic. Pick your most important pages and actually maintain them. A refresh date the engine can see is worth more than a fifth backlink.
Do not block the crawlers you want citing you
This one catches people. If your robots.txt blocks the AI crawlers, you cannot be cited by the engines those crawlers feed. Check that you are allowing the ones you actually want: GPTBot, ClaudeBot, and PerplexityBot, among others. Decide your posture on purpose, then make robots.txt agree with it. Blocking a crawler by accident is the most avoidable way to disappear from an answer.
So which job are you actually losing?
Most teams I talk to are doing the first job and ignoring the second. They rank fine. They have no idea whether they show up in the AI answer, or whether the answer represents them correctly. That gap is the whole problem, because rank alone now hides as much as it reveals.
So measure both. Keep the rank tracker you already have, and add a check for whether you show up in the synthesized answer and whether it gets you right. I built a free AI content visibility checker to make that less of a guessing game.
The bottom line is that AI Overviews did not replace classic SEO so much as stack a second layer on top of it. You still have to rank, since the pages that get cited are mostly the pages that already rank, but ranking on its own no longer guarantees the click. The work now is to win classic SEO so you are in the answer at all, and then format the page so the answer is easy to extract and quote.
If you want this from the practitioner side, with the formatting debates and the off-site trust angle, I took notes from a panel on AI search and GEO that is worth your time: notes from a panel on AI search and GEO.
Frequently asked questions
Do AI Overviews replace SEO?
No. AI Overviews pull from pages Google already trusts, so the citation pool is mostly the ranking pool. Classic SEO is still how you get into the answer. What changed is that ranking no longer guarantees the click. AI Overviews added a second job, getting cited, on top of the first one, ranking.
How do I get cited in an AI Overview?
Rank first, then format for extraction. Lead each page with a quotable answer in the first 40 to 60 words. Use question-shaped headings that mirror People Also Ask. Add comparison tables and schema. Keep cornerstone pages fresh, and make sure robots.txt allows the AI crawlers you want citing you.
What is the difference between SEO and GEO?
SEO is the job of ranking on the page. GEO, generative engine optimization, is the job of getting named inside the AI-generated answer. Right now they are the same motion, because AI answers cite pages that already rank. The difference is the outcome you measure: position for SEO, citation for GEO.
Do AI Overviews hurt traffic?
They lower clicks on the average query. When an AI Overview appears, the click rate on an organic result drops about 61%, per Seer Interactive across 2.43 billion impressions, and about 68% of searches now end with zero clicks. But cited brands earn around 120% more clicks per impression, so the traffic concentrates rather than vanishes.
Does ranking still matter if AI answers the question?
Yes. AI Overviews synthesize from pages that already rank, so if you cannot reach page one you almost never get cited. Ranking gets you considered, and getting named in the answer is what earns the click, so you need both. That is why classic SEO stayed necessary even though it stopped being sufficient on its own.
Sources
-
AI Overviews appeared on roughly 48% of tracked queries by February 2026. BrightEdge, 2026. ↩
-
Across 53 brands, 5.47M queries and 2.43 billion organic impressions, organic click-through fell about 61% on AI Overview queries, and brands cited inside the Overview earned about 120% more clicks per impression than uncited ones. Seer Interactive, AIO Impact on CTR: 2026 Update. ↩ ↩2
-
About 68% of US Google searches ended without a click in early 2026, per SparkToro’s analysis of Similarweb clickstream data. Search Engine Land, 2026. ↩
Working on this same shift?
I write about SEO, GEO, and getting found by AI search.
If this resonated, I'd love to compare notes.