SEO’s core disciplines like technical health, EEAT, and content quality still determine AI visibility. What’s new is how success is measured, how content must be structured for extraction, and how authority is assessed.
Open LinkedIn on any given morning and you’ll find two camps saying opposite things with equal confidence about the current state and future of SEO.
- One says nothing has really changed. Do good SEO, publish helpful content, keep your technical house in order, and AI visibility takes care of itself.
- The other says search as we knew it is finished. Rankings are a vanity metric, blue links are dying, and if you’re not doing AEO right now, you’re optimizing for a channel that won’t exist in eighteen months.
Let’s call them the Purists and the Prophets. Both are loud and come armed with a folder of case studies. And both are half right, which is exactly why the argument never resolves.
I’ve spent the last six years working with SEO and content teams, and my position sits squarely in the middle. SEO fundamentals are still indispensable. They’re the substrate everything else runs on. AEO is a real, necessary layer on top of that foundation, and pretending otherwise costs you visibility.
As search keeps transitioning to being entirely based on answer generation, here’s what has changed, what hasn’t, and what the future potentially holds.
Key Takeaways
- SEO fundamentals aren’t obsolete. Technical health, site architecture, EEAT, and quality content remain the foundation AI systems use to decide who’s worth citing. Google says this outright in its own generative AI guidance.
- Three things genuinely changed: how visibility is measured (citations and mentions, not just rankings and clicks), how content must be structured (retrievable, self-contained passages rather than page-length narratives), and how authority is assessed (brand consistency across the web, alongside backlinks).
- AEO is an additional layer, not a replacement discipline. You cannot be cited by a system that never crawled, indexed, or trusted you.
- Google Search and third-party AI assistants behave very differently. Advice that works for AI Overviews often doesn’t transfer to ChatGPT or Perplexity, and vice versa. Most of the confusion in this debate comes from people generalizing from one to all.
- The teams winning right now aren’t picking a side. They’re running SEO fundamentals and optimizing for AI visibility at the same time, with measurement that covers both.
Why Both Extremes — “SEO Hasn’t Changed at All” and “SEO Is Dead” — Get AI Search Wrong
Both extremes are reacting to something real, which is why the argument won’t die. The “nothing’s changed” camp is right that fundamentals still drive AI visibility. The “SEO is dead” camp is right that measurement and structure have shifted. But neither is seeing the whole picture.
What the “SEO Hasn’t Changed at All” Camp Gets Right
Sites cited in AI answers are, overwhelmingly, sites that were already doing SEO well, at least inside Google. Crawlability, clean structure, and credibility signals determine whether a system can find, parse, and trust your content in the first place.
The data backs this up. seoClarity’s analysis of 432,000 keywords found that 97% of AI Overviews cite at least one source from the top 20 organic results, and that position-one pages appear in AI Overviews more than half the time. So, while not a guarantee, ranking well is a strong signal.
Google is unusually direct about this in its own documentation. Its official guide to optimizing for generative AI features opens by confirming that SEO best practices remain relevant, because Google’s generative AI features are rooted in its core Search ranking and quality systems.
Google also addresses the terminology debate head-on, stating that from its perspective, optimizing for generative AI search is optimizing for the search experience, and therefore still SEO. That’s about as close to an official ruling on the Purist-versus-Prophet argument as we’re going to get.
What the “SEO Hasn’t Changed at All” Camp Gets Wrong
SEO purists defend the idea that we shouldn’t change SEO practices and strategy. But how can we keep doing the same when the analytics have changed so dramatically? Click-through rates have dropped significantly on queries where AI answers appear, which means you can hold your rankings and still watch your traffic fall.
Seer Interactive’s study of 3,119 informational queries across 42 organizations found organic CTR fell 61% on queries featuring AI Overviews between mid-2024 and September 2025, with paid CTR down 68%.
The more revealing number is what happened on queries without AI Overviews. Organic CTR still fell 41%, suggesting people are simply clicking less across the board.
Crucially, Seer found that sites cited within AI Overviews saw 35% higher organic CTR and 91% higher paid CTR. Being in the answer is now worth measurably more than merely ranking near it.
🎯 If your reporting stops at traditional rankings and organic sessions, you have no way to see that distinction — you just see decline.
One caveat worth mentioning: Seer’s 2026 follow-up found the decline has slowed, with AIO organic CTR rebounding in early 2026 and disrupting the assumption that further decline was inevitable.
Their own takeaway is that aggregate benchmarks can mislead, and your own data is the real strategic input. That’s good advice regardless of which camp you started in.
There’s also a blind spot at the content structure level. Classic SEO rewarded comprehensive pages that built an argument across several thousand words. AI retrieval works at the passage level. A page that only makes sense read start to finish is likely to be misquoted or skipped.
What the “SEO Is Dead” Camp Gets Right
The disruption is real and measurable. Generative-AI referral traffic grew more than tenfold in under a year, and, critically, citation behavior outside Google looks almost nothing like the organic SERP.
Adobe’s research found that generative-AI referral traffic to U.S. sites grew more than tenfold between July 2024 and February 2025, with those visitors browsing roughly 12% more pages per visit and bouncing 23% less often than non-AI referrals.
But the bigger challenge to SEO purist thinking comes from Ahrefs. Analyzing 15,000 long-tail queries, they found that only about 12% of citations from ChatGPT, Gemini, and Copilot also rank in Google’s top 10 for the same prompt.
Perplexity was the outlier at 28.6%, but for the rest, more than 80% of citations came from pages that don’t rank at all for the target query.
💡 Ahrefs’ own summary of the split is the single most useful sentence in this debate. AI Overviews follow the SERPs; AI assistants don’t.
What the “SEO Is Dead” Camp Gets Wrong
AI systems don’t generate answers from nothing. They retrieve and synthesize content that had to be crawled, indexed, structured, and trusted first. Declaring rankings irrelevant ignores that every citation still depends on infrastructure that is, definitionally, SEO.
Google’s documentation makes the dependency explicit. To appear in generative AI features, a page must be indexed and eligible to appear in Search with a snippet. There is no AEO shortcut around indexation. When someone tells you rankings are dead and only citations matter, ask them how they think a model finds the page it’s about to cite.
The “SEO is dead” framing also tends to arrive attached to a product. This isn’t a cheap shot, by the way. It’s observable in the search results for this very topic. Several of the highest-ranking articles arguing that traditional SEO is finished are published by companies selling AI visibility platforms, and each one closes with a demo request.
A discipline being declared obsolete by the vendors selling its replacement is a very old pattern in this industry, and it’s worth keeping in mind.
What Changes for SEOs in the AI Search Era
Six things have changed:
- How success is measured
- What counts as a click
- How content must be structured
- How queries get interpreted
- What counts as authority
- Which technical basics matter
Here’s each one, before and after.
From Rankings to Citations and Mentions
The core metric is changing from where you rank to how often you appear in the answer. That means tracking citations, mentions, and share-of-voice vs. competitors within each AI model alongside (not instead of!) the rankings you already monitor.
- Citations are when you’re linked as the source.
- Mentions are when your brand comes up in the answer without a link back to you.
- Share-of-voice (SoV) is how often you appear versus competitors for the prompts that matter to your business.
The Seer finding above gives this a hard number: cited pages earned 35% higher organic CTR than uncited ones on the same kind of query. That gap is the business case for tracking citations at all.
None of this replaces rank tracking. If you drop it entirely you lose your early-warning system for technical problems, because visibility decreases still show up there first.
For Google specifically, the Generative AI performance report in Search Console now shows how your content performs in AI features directly. Start there before buying a third-party tool.
From Clicks to Zero-Click Search
Fewer people click, and the drop is steep. 61% on AI Overview queries and 41% even on queries without them, per Seer.
What’s less discussed is that the clicks you still get tend to be better qualified, since anyone visiting after reading a full AI summary is already informed and further along.
Adobe’s data supports that read — AI-referred visitors browsed 12% more pages and bounced 23% less than other referrals.
The practical implication is that raw session volume is becoming a worse proxy for SEO’s business value. Conversion rate per session, assisted conversions, and branded search volume are all better indicators now.
From Full Pages to Retrievable Passages
AI systems retrieve passages rather than whole pages, so sections that only make sense in sequence get skipped or misread.
But this is also where the AEO industry has gotten furthest ahead of the evidence, and it’s worth being precise about what’s actually established.
Google’s July 2026 guidance explicitly tells site owners they don’t need to break content into small pieces for AI, saying its systems can already understand multiple topics on a page and surface the relevant part.
The same guidance says structured data isn’t required for generative AI search, and that perfectly semantic HTML isn’t necessary either. If you’ve been sold “chunking” as a Google ranking tactic, Google disagrees in writing.
What is supported by evidence is subtler and more durable. The Princeton GEO study (Aggarwal et al., KDD 2024) tested nine content modifications across roughly 10,000 queries and found that adding statistics, quotations, and source citations produced the strongest gains, with statistics improving visibility by around 41% on their primary metric. Notably, keyword stuffing was among the weakest approaches and could actively reduce visibility.
So the honest version of this advice isn’t “chunk your content for the robots.” It’s: write sections that stand on their own, answer the question in the heading immediately, and back claims with specific, verifiable evidence.
That serves human readers, it’s what the research actually supports, and it doesn’t depend on any particular platform’s current behavior.
Here’s a summary of the best on-page practices in 2026:
- Every H2 and H3 should state a specific question or claim, not a clever teaser.
- The first two or three sentences under each heading should answer it completely.
- Support claims with concrete figures and named sources rather than general assertions.
- Format comparisons and processes as tables or lists when that genuinely helps a reader.
From Keyword Matching to Intent and Entity Understanding
Modern AI search rarely answers your literal query. It expands it into a set of related sub-questions, gathers sources for each, and synthesizes across them. You’re no longer optimizing a page for a keyword; you’re building a defensible answer to a whole topic.
This is called query fan-out, and Google documents it plainly.
Its own example: the query “how to fix a lawn that’s full of weeds” fans out into sub-queries about herbicide selection, chemical-free weed removal, and weed prevention.
💡 Ahrefs’ research suggests the same mechanism explains why AI assistant citations diverge so far from the SERP — a page ranking sixth for several related queries can beat a page ranking first for only one.
One important caveat, again from Google’s documentation: creating separate pages for every possible query variation in order to capture fan-out traffic can trip its scaled content abuse policy. Cover your topic with genuine depth. Don’t manufacture a page per permutation.
From Backlinks to Brand Consistency Signals
Links still count, but AI systems weigh whether your brand is described consistently everywhere it appears. Research across AI platforms suggests brand search demand and entity recognition now predict citation frequency more strongly than backlink volume does.
The Princeton research points in the same direction from a different angle. Its most striking finding was an equalizer effect. Pages sitting at position five — normally low-visibility in AI answers — saw a 115.1% visibility increase when proper attribution was added. Credibility signals can partially substitute for rank.
The practical consequence is organizational as much as tactical. Consistent naming, accurate third-party listings, and real editorial coverage now sit inside the SEO remit alongside digital PR and brand management.
It’s worth noting the limit that Google draws, though. Manufacturing mentions across the web isn’t the shortcut it appears to be, and its guidance calls out inauthentic mention-seeking directly.
What Actually Matters Technically
Schema markup and semantic HTML are useful, but treat them as good practice rather than a magic AI unlock. Google states plainly that neither is required for generative AI visibility, though structured data still earns you rich results in classic Search.
The technical requirement that genuinely hinders AI visibility is simpler and older. Your page must be crawlable, indexable, and eligible for a snippet. Everything downstream depends on that.
If your site is coasting on blocked resources, broken canonicals, or JavaScript that hides content from crawlers, that debt has a direct visibility cost, and no amount of AEO tooling will route around it.
What Doesn’t Change in SEO in the AI Era
Six things haven’t changed at all, and they get far less attention because they’re less exciting to post about. These are technical health, site architecture, EEAT, keyword research, content quality, and genuine first-hand experience.
Technical SEO Health
In other words, crawlability, indexability, site speed, and mobile usability.
AI systems still have to reach and parse your site. Start with a full content inventory if you don’t already have one. You can’t fix what you haven’t mapped.
A page that’s slow, blocked, or broken is invisible to answer engines for exactly the same reasons it was invisible to Google in 2015, and per Google’s own requirements, an unindexed page is ineligible for AI features entirely.
Site Architecture and Internal Linking
A logical structure with meaningful internal links is crucial when building your SEO strategy and now also important for a successful AEO strategy.
It helps retrieval systems understand how your content relates to itself. Orphaned pages and flat, contextless architecture are as damaging as ever.
EEAT
EEAT didn’t get superseded by AI-era authority signals. It got more important.
The Princeton research found that content with clear attribution and evidence gets cited substantially more often, and Google’s guidance emphasizes first-hand expertise as a differentiator its AI systems actively look for.
Keyword and Intent Research
You still need to know what people are asking and why. The targeting logic evolved toward topics and question clusters rather than exact-match strings, but the underlying research discipline is unchanged.
Quality, Well-Researched Content
High-quality content is the throughline. AI systems reward accuracy, depth, and clarity, which are the same qualities good SEO always demanded, now enforced more strictly because a model that can’t verify a claim tends to route around it.
Human Storytelling and First-Hand Experience
This is worth more now than it was three years ago, and Google says so explicitly. Its guidance distinguishes commodity content — the generic “7 tips” article anyone could write — from non-commodity content built on genuine expertise, and states that a unique point of view based on personal experience is what stands out to its AI systems.
Original data, real client outcomes, and an actual opinion are the things generative models structurally cannot manufacture.
Traditional SEO vs. AEO Compared
Six dimensions separate the two disciplines in practice — goal, success metrics, content focus, interface, user behavior, and tooling. The table below lays them out side by side as a quick reference.
| Dimension | Traditional SEO | AEO / GEO |
|---|---|---|
| Primary goal | Rank in the SERP | Be cited and accurately represented in AI-generated answers |
| Success metric | Rankings, clicks, organic sessions | Citation rate, inclusion rate, AI referral traffic |
| Content focus | Keyword targeting, backlink authority | Evidence density, self-contained passages, entity clarity |
| Search interface | Ranked list of blue links | Synthesized, conversational answers |
| User behavior | Clicks through to explore | Receives a summarized answer first, clicks only if motivated |
| Example tools | Semrush, Ahrefs, Google Search Console | Profound, Ahrefs Brand Radar, Semrush AI Toolkit, GSC Generative AI report |
The Future of SEO With AI
Nobody knows for certain, but three trends are already visible and likely to intensify: AI agents acting on users’ behalf, brand consistency becoming a bigger trust signal, and measurement slowly catching up to reality. None of this means SEO is ending.
Agentic Traffic
AI agents that book, compare, and transact on users’ behalf are no longer hypothetical. Google’s own documentation now includes a section on agentic experiences, pointing site owners to agent-friendly design guidance and noting emerging protocols like the Universal Commerce Protocol.
Browser agents interact with sites by reading the DOM and the accessibility tree, which makes accessibility work an AI visibility issue.
Brand Consistency as a Ranking-Adjacent Signal
As models weight confidence in what they know about you, the gap between well-managed and sloppily-managed brand presence will widen.
Measurement Is Catching Up (Slowly)
Current tooling systematically undercounts AI traffic, though it’s improving. GA4 added a native AI Assistant channel in May 2026 that auto-captures traffic from ChatGPT, Gemini, and Claude. This is genuine progress.
But Perplexity still lands in Referral, and Google’s own AI Overviews and AI Mode clicks still bucket into Organic Search with no separate label, since the referrer looks identical to a regular Google click.
Search Console did launch a dedicated Generative AI performance report in June 2026, but it’s impressions only, no clicks or CTR yet, and still in limited rollout, so it doesn’t close this particular gap.
Referrer data also gets stripped when links open inside mobile app browsers, when users copy and paste URLs from an answer, and when redirect chains pass through HTTP, with estimates that 30–50% of AI-driven sessions land in GA4 as Direct with no source attribution at all.
Treat your AI referral numbers as a floor, not a total, and layer a custom channel group on top of the native one to catch what it still misses.
Conclusion: SEO and AEO Are Different Visibility Layers, Not Rivals
The SEO purists are right that fundamentals still carry the weight. Google’s own documentation says so. The prophets are right that the analytics have changed deeply and that AI assistants cite very differently from the SERP. Neither camp describes the whole reality, which is why picking one side leaves visibility on the table.
The version that works in practice is to keep the foundation in excellent condition, add the layer deliberately on top of it, and measure both.
If your site needs a complete audit, from SEO fundamentals to the AEO layer, know that that’s the work I do. Happy to take a look.
FAQs
How is AI changing SEO in 2026?
The biggest changes are in measurement, structure, and authority. Success is increasingly defined by whether you’re cited in AI-generated answers rather than only where you rank. Seer Interactive found sites cited in AI Overviews earn 35% higher organic CTR than those that aren’t. Content needs to work at the passage level, and consistent brand representation now sits alongside backlinks as an authority signal. The underlying work, including technical health, quality content, and genuine expertise, is largely unchanged.
Can SEO be replaced by AI?
No, though the job description keeps widening. AI systems generate answers by retrieving and synthesizing content that was crawled, indexed, and judged trustworthy first. Every one of those steps is SEO. Google’s own guidance states that its generative AI features are built on its core Search ranking systems, and that a page must be indexed and snippet-eligible to appear in them at all. What AI changes is the interface and the scoreboard, not the requirement that someone make content findable, parseable, and credible.
Which AI tool is best for SEO?
There isn’t a single best one, and the category is changing too fast for any answer to stay accurate for long. You’ll want traditional coverage (Semrush or Ahrefs and GSC) plus something that tracks AI visibility and citations. Start with the free option. Search Console’s Generative AI performance report shows how your content performs in Google’s AI features directly. Google also cautions against third-party tools claiming access to internal ranking metrics — no external tool has it.
How do you stay current with SEO and AI search changes?
I stay current with changes in SEO and AI search mostly through a handful of sources I check consistently rather than trying to read everything. I subscribe to newsletters from Ahrefs, Search Engine Land, and Search Engine Journal (my favorite of the three). I watch the Grow and Convert Marketing Show on YouTube. And I read Google’s official documentation directly and watch Google Search Central videos.
What is SEO for AI search called?
You’ll see three terms used, often interchangeably and not always carefully. SEO is optimization for search engines and rankings. AEO, short for Answer Engine Optimization, is optimization for being the cited answer when a system responds directly. GEO, or Generative Engine Optimization, is optimization for being referenced or synthesized within AI-written responses. Academic researchers note the terminology remains contested. The more useful framing is that these are overlapping layers running on the same technical foundation, not three competing disciplines fighting over your budget. (For the fuller breakdown, including where AIO and LLMO fit in, read my AEO vs. GEO vs. AIO vs. LLMO guide.)
