SEO for AI Search in 2026

Search changed more in the last two years than in the previous ten, and pretending otherwise is expensive. SEO for AI search is now a different job than it was in 2023. Google's AI Overviews answer the question right on the results page, so people click less. Pew Research Center tracked real browsing and found users clicked a traditional link 8% of the time when an AI summary appeared, against 15% when it did not. That is close to half the clicks, gone. The good news is that most of what makes content genuinely good still wins. You just have to understand the new rules and stop fighting the old game.

Search stopped meaning clicks

For twenty years the deal was simple. Rank well, earn the click, get the visit. AI Overviews broke that deal for a big slice of queries by answering on the page. The data is not subtle. In Pew's study, only 1% of people clicked a link inside the AI summary, and 26% ended their session entirely after seeing one, against 16% without. You can read the Pew findings in full. Zero click behaviour was already climbing. Similarweb put it at 56% of searches in May 2024 and 69% a year later.

This does not mean the web is over. It means informational, top of funnel queries, the "what is" and "how do I" searches, now often get answered without a visit. About 18% of searches triggered an AI Overview in Pew's window, and question phrased searches triggered one around 60% of the time. If your traffic leaned on exactly those queries, you have already felt the floor move.

Your content now has two jobs

Here is the mental shift. Content used to have one job: rank high enough to earn the click. Now it has two. It still needs to rank, and it also needs to be the source the AI quotes when it writes the answer. Those are related but not identical, and optimising for both is the whole game in 2026.

Getting quoted matters more than it first sounds, because a citation in an AI answer puts your brand in front of the user even when they do not click, and when they do click, they arrive warmer. Seer's analysis found brands cited in AI Overviews earned about 35% more organic clicks than uncited ones, and that AI referred visitors converted several times better than traditional organic. Visibility inside the answer is the new front page.

Answer engine optimization, in practice

Optimising to be quoted, sometimes called answer engine optimization, is mostly about making your content easy for a machine to read, trust, and lift a clean answer from. In practice that looks like a few habits.

• Answer the question directly, early. State the answer in a sentence or two near the top, then expand. AI summaries in Pew's data ran a median of 67 words, so a crisp, self contained answer is what gets pulled.
• Structure for extraction. Clear headings phrased as real questions, short paragraphs, and lists where they genuinely fit. Make the shape of the page obvious at a glance.
• Add structured data. Schema for articles, FAQs, products, and how-tos helps machines understand what each part of your page actually is.
• Cover the follow ups. Answer the obvious next questions on the same page, so you are the complete source, not a fragment.

None of this is a trick. It is just clarity, and clarity has always helped humans too. The only difference now is that a machine reads first and decides whether you are worth quoting.

Why thin content lost, in Google's own words

If Google has rejected your content as low value, this is the section that matters. The temptation in an AI world is to mass produce articles cheaply. That is precisely what Google moved to stop. Its scaled content abuse policy is blunt about it. In Google's own words, scaled content abuse is when many pages are generated mainly to manipulate rankings and not to help users, and it applies no matter how the content is created, by automation, by people, or a mix. You can read the spam policies and the policy announcement directly.

Read that carefully, because it is actually liberating. Google is not against AI assisted writing. It is against unoriginal, low value pages produced at scale to game search. The fix is not to write less with AI. It is to make every page genuinely worth reading. That is where experience, expertise, authoritativeness, and trust come in, the E-E-A-T standards Google leans on. Real author identity, first hand experience, original data or examples, and plain accuracy. A leaked look at Google's ranking systems even revealed an internal content effort signal, an estimate of how much real work a page represents. Thin content scores badly on exactly that. This article, and this whole site, is written to that bar, which is the only approach we would ever recommend to a client.

Most of SEO did not actually change

Amid the noise, it is worth saying plainly: the fundamentals still work. Understanding what a searcher actually wants and giving it to them better than anyone else. Genuine depth on topics you know. A fast, clean, mobile friendly site. Sensible internal links so both people and crawlers can navigate. Earning links and mentions by being worth citing. These moved the needle in 2015 and they move it now.

What changed is the reward at the top and the penalty at the bottom. Ahrefs found the number one result can lose up to 58% of its clicks when an AI Overview sits above it, so ranking first is worth less than it was. And thin content is now actively demoted, so the floor is lower too. The middle path, real quality that earns both rankings and citations, is the safe one. It is also, not coincidentally, the approach behind properties we run like brandligo.com.

How to actually get cited by AI

To be the source an AI quotes, be the source worth quoting. A few things reliably help. Be a primary source, publishing original data, first hand tests, or a genuinely useful synthesis rather than a rehash of the top ten results. Be quotable, with clean, factual sentences a model can lift without ambiguity. And build real authority in your niche, since models lean on sources they, and the wider web, treat as trusted.

It also helps to know where AI answers tend to draw from. Pew found Wikipedia, YouTube, and Reddit among the most cited sources. You will not out rank Wikipedia, but you can be the specialist, authoritative source for your specific corner, which is exactly what a travel guide like traventigo.com aims to be for its topics. Depth and originality in a narrow lane beat shallow breadth every time.

Measure what matters now

If you only track clicks, AI search will look like pure loss and you will make bad decisions off the back of it. Widen the lens. Watch impressions and rankings, which show you are still surfacing even when clicks dip. Track whether you appear in AI answers for your key queries. Watch branded search, since AI visibility often lifts the number of people searching your name directly. And above all, follow conversions and revenue, not just sessions.

That last point is the real reframe. If AI referred visitors convert several times better than classic organic, as Seer's data suggests, then fewer, warmer visitors can be worth more than a flood of cold ones. Traffic was always a proxy for value. In the AI era, measure the outcome you actually care about, which is the same discipline we bring to every growth engagement.

A playbook you can start this month

You do not need to panic, and you do not need a rebuild. Start here. Audit which of your pages depend most on informational queries exposed to AI Overviews, and shore up their depth and structure. Rewrite your best pages to answer their core question in the first breath, then go deeper than anyone else does. Add or tidy your schema. Strengthen author identity and first hand experience across the site. And shift your reporting from clicks alone to citations, branded search, and conversions.

Search is not dying. It is raising its standards. The sites that win from here are the ones with real expertise, presented clearly enough for both a person and a machine to trust. If you want a clear eyed audit of where you stand and what to fix first, that is a good first conversation to have with us.

Frequently asked questions