AI Overviews vs Featured Snippets: What Changed

AI Overviews vs featured snippets comes down to one shift: the AI Overview is the new position zero, it sits above the featured snippet, and it synthesizes an answer from many sources instead of lifting one passage from one page. A featured snippet copies a single winner word for word. An AI Overview blends several pages into a fresh paragraph and cites all of them. So one source became many, one slot moved down, and the clicks the snippet used to harvest are quietly leaking into the box on top of it.

That is the whole story in a paragraph. The rest of this is the machinery: where each one sits, who gets cited, what the click-through data actually says, and what you should optimize for now that the stack reshuffled.

One honest note before we start. I am an engineer, not a hype guy, and this stuff moves weekly. Google ships changes to these boxes constantly, the percentages drift, and anyone who claims a perfect playbook is selling you a course. I will give you the solid parts and flag the wobbly ones as we go.

Key takeaways

• The AI Overview is the new position zero. According to The HOTH, an AI Overview sits at position zero above the featured snippet and synthesizes multiple sources, while a snippet pulls from a single page.
• Snippets still pull strong CTR, when they show. Seer Interactive's 2025 data found the cited URL in a featured snippet earns roughly 42.9% click-through rate, while an AI Overview citation adds about +1.08 percentage points.
• Most searches end with no click. Zero-click searches now run roughly 55% to 65%, per The HOTH, and featured snippets account for about 8% of all SERP clicks.
• Google says AI clicks still count. Google's documentation states clicks from AI features count as normal clicks in Search Console and tend to be higher-quality engagement.
• Structure gets you lifted. Contently found tables get extracted 81% of the time versus 23% for prose, which is why the same formatting wins both surfaces.

The core difference, in one paragraph

A featured snippet is Google quoting one page. An AI Overview is Google writing its own answer from several pages and telling you which ones it used. That is the entire distinction, and almost everything else follows from it. The snippet is extraction: Google finds the single best passage that answers your query and shows it verbatim, with a link to the page it came from. The Overview is synthesis: Google's Gemini model reads a cluster of pages, drafts a fresh summary in its own words, and attaches a handful of cited links. One copies a winner. The other blends a committee.

That difference changes the whole game for you. In the snippet world, one page won and everyone else lost the box. In the Overview world, three, five, or more sites can get named in the same answer, which means more seats at the table but a fuzzier definition of "winning." For the full anatomy of the AI box itself, our explainer on what Google AI Overviews are breaks down how the summary gets built and who it cites.

AI Overview vs featured snippet, side by side

Here is the clean comparison. I find a table settles more arguments than three paragraphs, and the engines apparently agree, which is a theme we will come back to.

Read down the "how content is chosen" row and you see the real fork. A snippet is a retrieval-and-copy operation: deterministic, repeatable, the same passage until something changes. An Overview is retrieval-and-generation: it pulls candidate pages, then a language model writes something new each time. That generation step is why your brand can show up in the box on Monday and vanish on Tuesday without you touching a thing. If you want the engineering view of that candidate-selection step, we wrote up how AI engines choose sources in detail.

The SERP stack in 2026: AIO, then snippet, then organic

The 2026 results page is a layer cake, and knowing the order tells you where the attention goes. When Google decides a query deserves an AI answer, the AI Overview lands first, at the top, owning the screen before anything else loads. Below it, if Google also serves one, sits the featured snippet. Below that, the familiar organic blue links. Then ads, people-also-ask, the usual furniture. The HOTH lays out exactly this stacking: AI Overview on top, snippet under it, organic beneath.

So the thing old SEO folks called position zero, the featured snippet, got demoted to roughly position one-and-a-half. The new position zero is the AI box. This is not a rename for fun. Position on the page is attention, and attention is the whole currency here. The slot that sits first gets read first, and on a phone screen it often gets read instead of everything below it. Our Google AI Mode and SEO guide covers where this stack is heading next, because AI Mode is the conversational version of the same takeover.

Not every query gets an Overview

Important caveat, and it is good news for snippet fans. Google does not slap an AI Overview on every search. Plenty of queries still return a featured snippet with no Overview at all, especially short factual lookups, navigational queries, and anything where a one-line answer obviously does the job. So the snippet did not get deleted from the page. It just got a roommate who showed up on the more lucrative, higher-intent questions and started hogging the couch. When the Overview is absent, the snippet is back to being the top dog.

What actually happened to snippet traffic and CTR

Here is the part everyone wants and most articles fumble, so let me be careful with the numbers. Featured snippets still convert well when they appear. According to Seer Interactive's 2025 data, cited by DBS, the cited URL in a featured snippet pulls roughly 42.9% click-through rate. That is a genuinely strong number. Nearly half of people who see that snippet click the source. The snippet was never a traffic killer on its own. It answered the easy stuff and passed the meatier intent down to a click.

The AI Overview is a different animal. Seer's data via The HOTH found that being cited in an AI Overview adds only about +1.08 percentage points of organic CTR. Read that next to the snippet number and the contrast is stark: a snippet citation is a big click magnet, an Overview citation is a small click nudge. The Overview is not mainly a click machine. It is a brand-visibility machine. You get named in front of everyone who reads the answer, but most of them do not click anybody, including you.

And that is the real story behind the traffic worry. Featured snippets account for roughly 8% of all SERP clicks, per the figures The HOTH and DBS cite, so they were always a slice, not the whole pie. The bigger force is zero-click behavior. Zero-click searches now run roughly 55% to 65% according to The HOTH. More than half the time, the search ends on the results page. The AI Overview did not invent zero-click. It poured gasoline on a fire that featured snippets and knowledge panels lit years ago.

The click math, made honest

Let me put the trade-off in plain arithmetic so nobody oversells it to you. If you win a featured snippet on a query that does not trigger an Overview, you get the slot plus that fat ~42.9% CTR. Good day. If that same query starts triggering an AI Overview, the snippet may still appear, but it now sits under the box, and a chunk of readers get their answer up top and never scroll to it. Your snippet CTR on that query drops, not because your page got worse, but because the box above you intercepted the attention. You did nothing wrong. The page just got a new ceiling. This is why measuring per-query, over time, matters more than ever, and we will get to that.

Do featured snippets still matter? Yes, with an asterisk

Featured snippets still matter, and the asterisk is "when they show up and when you can own them." Three reasons they are not dead. First, the CTR is still excellent. A ~42.9% click rate on the cited URL is a number most pages would trade a kidney for. When the snippet appears and you hold it, you win real clicks, not just a brand mention. Second, plenty of queries still serve a snippet with no Overview at all, so the slot has not vanished, it just shares the page sometimes.

Third, and this is the part people miss, the work that wins a featured snippet is the same work that wins an AI Overview citation. A clean, concise, self-contained answer near the top of your page is exactly what the snippet extractor wants and exactly the kind of passage the Overview synthesizer likes to pull from. So optimizing for snippets is not a legacy tactic you do out of nostalgia. It is the on-ramp to the AI box. You are not choosing between the two. You are doing one piece of work that pays in both places. For the broader frame on why this convergence happened, our answer engine optimization guide and the what is AEO explainer both go deeper than I can here.

Where snippets quietly lose

I promised honesty, so here is the loss column. On high-value, high-intent queries, the ones with money behind them, Google is most likely to fire an AI Overview, which is precisely where it hurts most to lose attention. The snippet survives best on low-stakes factual queries and gets squeezed hardest on the commercial ones you actually care about. So "snippets still matter" is true, but the slots that matter most to revenue are exactly the slots most likely to sit under an Overview. Plan accordingly.

How to optimize for both at once

You optimize for both by writing one clean answer block and wrapping it in structure that machines can lift. The good news is there is no separate AI-Overview checklist that contradicts the snippet checklist. They want the same things. Here is the concrete version, the stuff I would actually put in a content spec.

Lead with a tight, self-contained answer

Put the direct answer to the page's main question in the first two or three sentences, written so it makes sense even when an engine rips it out and drops it in a box. Aim for roughly 40 to 55 words for the core answer, because that is the size both the snippet extractor and the Overview synthesizer like to grab. Do not bury your best sentence in the conclusion. Engines pull most of what they cite from the top of the page, so a buried answer is a wasted answer. This single habit moves the needle more than any schema trick.

Structure for extraction, not just for humans

Use question-style headers, short paragraphs, ordered and unordered lists, and at least one comparison table. This is not decoration. Contently found that tables get extracted 81% of the time versus just 23% for plain prose. That is not a rounding error, that is a 3.5x difference in how often your content gets pulled. So when you have data that fits a grid, put it in a grid. The engines genuinely prefer it, and so do skim-readers, which is everybody. Our AI content optimization piece has the full formatting checklist if you want the longer list.

Add schema, but keep your expectations calm

Mark up your pages with the obvious structured data: FAQPage, HowTo where it fits, Article, Organization. Schema helps machines understand what your content is and makes the FAQ and how-to formats easier to parse and lift. I will be straight with you, though: schema is a clarity aid, not a magic citation button. It improves your odds of being understood correctly. It does not force Google to cite you. Treat it as table stakes that you do once and maintain, not as the lever that fixes everything. Our schema markup for AI search guide covers which types actually earn their keep.

Be the kind of source an Overview wants to synthesize

This is the squishier one, and the most important. The Overview blends several pages, so you want to be the page that most cleanly answers the specific sub-question, with named facts it can trust. That means real statistics with sources, dated claims, named authors, and outside citations to credible references. Verifiability is an E-E-A-T signal that engines reward, which is the same reason this article keeps naming The HOTH and Seer and Contently instead of waving its hands. Be the page that is easy to quote and safe to trust, and you become the page that gets pulled into the synthesis.

None of this is exotic. It is being clearer and better organized than the next page, on purpose. That is a low bar that, in practice, almost nobody clears, which is exactly why the opening exists. If you want the engine-by-engine playbook beyond Google, the full GEO playbook and our how to appear in Google AI Overviews guide are the next two stops.

The honest bit: you control the content, not the box

Here is the trade-off nobody likes to say out loud. You control what you publish. You do not control whether Google decides to show an AI Overview on a given query, which sources it picks, or how it words the answer. You can write the cleanest, most liftable, most trustworthy page on the internet and still get left out of a particular box, because the box is generated and Google is steering it, not you. That is not a reason to give up. It is a reason to measure instead of guess.

Because the Overview wobbles run to run, a single manual check tells you almost nothing. You search once, see your name, feel great, and the next person gets a box that does not mention you at all. To know your real citation rate you have to run the same buyer questions repeatedly, across time, and average it out with some sense of the margin of error. That is repetitive volume work no human wants to do by hand, and doing it once a quarter on a whim is not measurement, it is vibes.

This is the part where AI Citation Monitor earns its keep. It runs your important questions across ChatGPT, Perplexity, Gemini, and Google AI Overviews on a schedule, then reports your citation rate with a confidence interval, so you know the number is real and not one lucky run, plus your share of voice against competitors and prescriptive fixes for the gaps. There is a free instant check if you just want to see where you stand right now. For the metric itself, our citation rate glossary entry defines it cleanly, and the AI citation tracking guide covers the full how-to.

The point is not to obsess over one figure on one day. It is to stop guessing about a moving target. Most brands have no real idea what Google's AI box says about them, so the ones who actually keep score get a quiet, slightly unfair head start.

So what actually changed, in one breath

AI Overviews vs featured snippets is the story of position zero changing hands. The snippet copies one page and sits below. The Overview synthesizes many pages and sits on top, as the new position zero, where it absorbs a chunk of the attention and clicks that snippets used to collect. Snippets still pay well when they appear, with that ~42.9% CTR per Seer, but they appear less often on the queries that matter, and zero-click behavior keeps climbing toward two-thirds of searches.

What you do about it has not really changed, which is the comforting part. Answer the question fast in a clean, self-contained block. Structure it with tables, lists, and a real FAQ, because structure gets extracted far more than prose. Earn trust with named facts and sources. Add sensible schema and keep calm about it. Then go measure whether Google is actually saying your name, because that is the one thing the content alone cannot tell you. When you are ready for the next layer, our what is AI search overview is where I would send you.

FAQ

What is the difference between AI Overviews and featured snippets?

A featured snippet lifts one passage from one page and shows it word for word. An AI Overview generates a brand new summary from several pages and cites all of them. The Overview sits above the snippet on the page, which is why people now call it the new position zero. The snippet copies one source, the Overview synthesizes many.

Which sits higher on the page, an AI Overview or a featured snippet?

The AI Overview sits higher. When Google shows an AI Overview, it lands at the very top of the results, above the featured snippet and above the organic blue links. The featured snippet used to be position zero, the highest spot. Now the Overview took that crown, so the snippet got bumped down a slot.

Do featured snippets still matter in 2026?

Yes, partly. According to Seer Interactive's 2025 data, the cited URL in a featured snippet still pulls roughly 42.9% click-through rate, which is far higher than an AI Overview citation adds. Snippets just appear less often now, since an AI Overview often replaces them. So they matter when you can win them, but you cannot count on the slot being there.

Did AI Overviews kill featured snippet traffic?

They dented it, not killed it. Featured snippets account for roughly 8% of all SERP clicks per the data cited by The HOTH and DBS, and zero-click searches now run about 55% to 65%, so more searches end without a click at all. When an AI Overview appears, it often eats the attention the snippet used to get. The snippet still drives clicks when it shows, just less often.

How do I optimize for both AI Overviews and featured snippets?

Answer the exact question in the first two or three sentences, in a clean self-contained block of about 40 to 55 words. Add structure like tables, lists, and a real FAQ, since Contently found tables get extracted 81% of the time versus 23% for prose. Then earn trust with named authors, dated stats, and outside citations. The same content habits win both surfaces.

Does Google count clicks from AI Overviews?

Yes. Google's own documentation says clicks coming from AI features count as normal clicks in Search Console, and it claims those clicks tend to be higher-quality, more engaged visits. So a citation in the AI Overview is not invisible to your reporting. You just have to accept that fewer people click overall, because most searches now end inside the box.