Most write-ups on AI Overviews and traffic decline stop at the obvious victims — blogs and news sites whose informational pages get answered in place. The more instructive case is Reddit. Google spent 2023 and 2024 elevating forum and community content, which sent Reddit's search visibility soaring; it signed a reported $60-million-a-year deal in February 2024 to train its AI on Reddit's posts; and Reddit became one of the single most-cited sources inside AI answers, including Google's own AI Overviews. By every measure that is supposed to protect a site from the AI-search shift, Reddit was maximally protected. And in 2026 its stock still fell hard after its CEO warned that Google's AI Overviews were making search referrals volatile — the platform Google boosted, licensed, and cites most is watching the same clicks get intercepted at the results page. That is the real lesson of SEO versus AI Overviews, and it is not "rank better" or even "go where Google is sending traffic now." It is that any distribution built on borrowed referrals — from Google, or from whichever platform Google is currently favoring — is structurally fragile, and the durable response is to stop being a tenant on someone else's discovery surface. This guide uses the Reddit case to show why, then lays out the distribution strategy that survives it.
The usual story about SEO versus AI Overviews goes like this: Google now answers many queries with an AI summary at the top of the results page, so a page can rank exactly where it always did and still lose most of its clicks, because the searcher got the answer without scrolling to a link. That is true, it is well measured, and it is covered in depth in AI Overviews are reducing organic clicks and the publisher traffic collapse. The obvious victims are informational blogs and news sites whose pages an AI can summarize away.
This guide takes a different route to the same conclusion, because the most instructive case in the whole shift is not a struggling blog. It is Reddit — the single site that, by every measure supposed to protect you from AI search, was maximally protected, and got hit anyway. Google spent 2023 and 2024 deliberately elevating forum and community content, which sent Reddit's search visibility soaring. Google signed a reported sixty-million-dollar-a-year deal in February 2024 to train its AI on Reddit's posts. Reddit became one of the most-cited sources inside AI answers, including Google's own AI Overviews. And in 2026 Reddit's stock still fell sharply after its CEO warned that Google's AI Overviews were making search referrals volatile. If the platform Google boosted, paid, and cites most cannot rely on Google's clicks, no one renting that channel can. The real lesson of SEO versus AI Overviews is not "rank better," and it is not even "chase whatever Google is boosting now." It is that borrowed distribution is fragile by construction — and that changes where you should put your effort.
To see why Reddit is the sharp case, you have to remember how much Google favored it. Around 2023, Google recognized that people were appending "reddit" to their searches to escape a web of SEO-optimized pages and reach real, first-hand opinions. Google leaned into that preference hard — surfacing forum and discussion content far more prominently — and Reddit's organic Google visibility climbed steeply through 2024. For a stretch, ranking advice half-jokingly reduced to "be Reddit," because Reddit was winning the exact queries independent sites used to.
Then Google put money behind the relationship. In February 2024, the same week Reddit filed to go public, the two signed a data-licensing agreement — reported at about sixty million dollars a year — giving Google access to Reddit's content to train its AI models and surface Reddit across Google products. On top of the traffic favor and the cash, Reddit's user-authored answers turned out to be ideal raw material for generative answers, so Reddit rose to become one of the most-cited domains across AI systems — by some 2026 citation indexes the single most-cited source across the major AI engines, and consistently a top source inside Google's AI Overviews themselves. Ranking favor, a paid data deal, and pole position in the AI citations: if any site was set up to be immune to the AI-search transition, it was this one.
It was not immune. In its 2026 quarterly results, Reddit reported strong business numbers — revenue up roughly 61% year over year — and its stock still dropped sharply, because the CEO warned that Google's AI Overviews were making search referral traffic volatile and "choppy," with more volatility expected. The market's read was blunt: no matter how well Reddit monetizes, a large share of its audience arrives through Google, and Google is increasingly answering the query itself instead of forwarding the click. Reddit's CEO made the point directly, arguing that the old "ten blue links" had created enormous value by sending people out to sites, and that an AI Overview summarizing a discussion is not the same as sending someone into the discussion — people who want Reddit usually want multiple human perspectives, not a machine's paraphrase of them.
The mechanism is identical to the one that hits any publisher: the AI answer sits above the links and resolves the query, so the click becomes optional and most people skip it. Pew Research measured the size of that effect — users clicked a result 8% of the time when an AI summary was present versus 15% when it was not, and clicked a link inside the summary only about 1% of the time. Similarweb put zero-click Google searches at 69%, up from 56% a year earlier, and AI Overviews now trigger on roughly half of all searches. Reddit is subject to exactly those numbers. Being cited inside the Overview does not save it, because a citation is not a visit — the whole design of the feature is that the user does not need to leave. Reddit's reported response has been telling: weighing whether to keep renewing the Google data deal at all, and, like other large publishers, considering how much access to keep granting a search engine that increasingly keeps the traffic. That is what it looks like when the favored tenant starts questioning the lease.
Strip the specifics away and the principle is simple. Reddit did not do anything wrong; it did everything a site is supposed to do to win in Google's world, and Google changed what winning in its world pays out. That is the defining property of a borrowed channel: the intermediary sets the terms, and can reset them whenever its own product strategy shifts. Search was always a rented channel — you did the SEO work and Google decided whether to forward an audience — and for two decades the rent was cheap and the traffic reliable. AI Overviews are Google quietly raising the rent to, often, "no click." The reason this is a distribution problem and not an SEO one is covered in SEO in the age of AI search: there is no ranking above the answer box, so no amount of optimization restores a visit the interface has decided to keep.
Here is the part most people miss, and it is why the Reddit case matters more than another blog-traffic chart. The tempting response to "Google is answering queries itself" is "then go where Google is sending traffic now" — pile into Reddit, or whatever platform Google is currently boosting, and ride its coattails. Reddit is the proof that this just relocates the fragility. You would have spent 2024 building a Reddit-dependent strategy on the strength of Google's favor, and 2026 watching that same favor become a liability the platform itself is trying to renegotiate. Chasing the currently-favored surface is still renting; you have just changed landlords. The only move that actually addresses the root is to stop letting any single intermediary's product decision determine your reach.
The distribution strategy that holds up does not try to win the borrowed channel back. It reduces how much of your reach depends on any borrowed channel at all, on three fronts that map directly onto the failure modes above.
If discovery is moving in-feed, be in the feed. Native content on the platforms themselves — short and long video, carousels, image posts, and text threads — is surfaced by each platform's own recommendation engine, with no external click required, so the "will they click through to my site" question that AI Overviews keep answering "no" to never gets asked. And the Reddit angle cuts the other way here too: instead of treating Reddit as a search-referral middleman you hope Google forwards visitors from, you can show up on Reddit as a genuine first-party presence, posting and engaging in the communities that matter to you. Building a brand presence on Reddit is a distribution channel you control your behavior in; being a page Google happens to rank via Reddit is not.
The most durable distribution passes through no algorithm at all. Email is the clearest case — a subscriber list is a direct line, and no AI Overview, ranking change, or platform data-deal renegotiation can sit between your newsletter and the inbox. It will never be your largest channel by raw reach, but it is the one that cannot be repriced out from under you, which is precisely the property Reddit just learned its Google traffic never had. Converting a share of your distribution from rented to owned is the single most defensive thing you can do.
The concentration effect that made Reddit a top AI citation is also available to you, at your scale. AI systems favor recognizable, authoritative, consistent entities, and users increasingly search for and ask about brands by name rather than typing generic queries. A coherent brand and a recognizable persona — the same voice, face, and look across everything you publish — makes you the kind of entity that gets named in an answer and sought out on purpose, which is the offensive counterpart to the defensive moves above. AI visibility beyond SEO and getting recommended in chat-driven discovery go deep on that work; the point here is that being nameable is a distribution asset in the AI era, not a vanity metric.
The strategy is sound and its execution is where it dies. Being present natively across many platforms, participating in communities, feeding an email list, and holding a consistent brand across all of it is several times the content the old model needed. Under search dependence you could write one strong page and let SEO distribute it; the post-Overview model asks you to produce video, carousels, posts, a blog, and a newsletter, sized and voiced for each surface, on a cadence, everywhere at once. That is not a strategy gap anymore — it is a production-capacity gap, and it is exactly where most brands stall and end up defaulting back to the one rented channel they know. It is the same wall described in building an automated social content engine: the winning distribution plan is only real if you can produce at the volume it demands without the quality collapsing.
Kompozy is built to remove that volume ceiling, which is the specific thing that keeps the sound strategy from being actionable. It is a full AI content generation and multi-platform publishing engine — 18 output formats across the eight social platforms plus blog and email — so instead of writing one page and hoping a borrowed channel forwards a click, you generate a coordinated native spread and push it straight onto the surfaces you occupy. One source idea becomes Persona Shorts and longer avatar video for the video feeds, carousels and photo posts for the image surfaces, platform-shaped text posts for the conversational feeds, a blog article for whatever search traffic remains, and — the un-disintermediable core of the whole plan — an email newsletter for the audience no algorithm gatekeeps.
The Reddit lesson maps onto two capabilities in particular. First, owned-audience distribution is a first-class output, not an afterthought: the newsletter is generated in the same pass as everything else, so the channel that cannot be repriced actually gets fed instead of being the thing you never get around to. Second, Kompozy can publish directly to communities including Reddit through its Direct Connect destinations, so Reddit becomes a surface you post to as a first-party presence rather than a search middleman you depend on — the exact reframe this guide argues for. And because brand nameability is what earns direct AI citation, consistency is enforced by the engine rather than left to willpower: a Persona Brief governs voice and banned words on every generation, Gemini face-lock keeps the persona's face consistent across images while HeyGen holds the same avatar in video, and HyperFrames renders brand-exact styling, so the same recognizable brand answers on every surface — which is what both audiences and AI systems reward. The strategy behind that identity work is in identity-first AI video.
Then Autopilot closes the loop the strategy depends on: it schedules and publishes that spread across the supported platforms plus email and blog from one queue, behind a per-post review gate so a human signs off before anything ships. That is the operational form of owning your distribution — your content lands directly in front of people on the platforms and inboxes you control, on a cadence, without waiting for a Google click that is increasingly withheld. The honest scope: Kompozy cannot make Google forward a click it has decided to keep, and it cannot force an AI Overview to cite you — Reddit's experience shows even maximal favor does not guarantee traffic. What it removes is the volume ceiling that makes the only durable response — presence everywhere plus an owned audience — impossible for a small team. If you depend on a handful of search-fed pages, classic SEO discipline is the cheaper call. Kompozy earns its place the moment you accept the Reddit lesson: that no borrowed channel is safe, and the way out is to build distribution you actually own, at a volume that used to require a newsroom. For the broader capability picture, see AI content engines for social media.
SEO versus AI Overviews is usually framed as a fight blogs are losing, but the clearest evidence is Reddit — the platform Google elevated in search, paid a reported sixty million dollars a year to license, and cites most inside its AI answers, whose stock still fell hard in 2026 when its CEO warned that AI Overviews were making search referrals volatile. If the favored, funded, most-cited tenant cannot count on Google's clicks, the takeaway is not "rank better" or "go where Google is sending traffic now." It is that borrowed distribution — from Google, or from whatever platform Google currently boosts — is fragile by design. The durable response is to stop being a tenant: publish natively where attention lives, participate directly in the communities that matter, feed an audience you own like email, and build a brand nameable enough to be cited on purpose. The only thing standing between that plan and most brands is the volume it takes, and volume is a solvable, systems-level problem.
Yes, and that is the surprising part. Reddit had every advantage against the AI-search shift: Google elevated forum content in 2023-2024, which sent Reddit's search visibility soaring; Google signed a reported $60-million-a-year deal in February 2024 to train its AI on Reddit's content; and Reddit became one of the most-cited sources inside AI answers, including AI Overviews. Even so, in 2026 Reddit's stock fell sharply after its CEO warned that Google's AI Overviews were making search referral traffic volatile. Being the platform Google favors and cites most did not stop the clicks from being intercepted at the results page.
Around 2023 Google shifted toward surfacing first-hand, community, and forum content — people were appending "reddit" to queries to escape SEO-optimized pages, and Google leaned into that. Reddit's Google search visibility rose steeply through 2024 as a result. Google also signed a data-licensing deal, reported at about $60 million a year in February 2024, to train its AI models on Reddit's posts. So Reddit's search traffic was a deliberate product decision by Google, which is exactly what makes it a borrowed channel Google can reprice.
The measurements converge on a large loss. Pew Research found people clicked a search result 8% of the time when an AI summary appeared, versus 15% when it did not — about half — and clicked a link inside the summary only around 1% of the time. Similarweb reported zero-click Google searches rising from 56% to 69% in a year. AI Overviews now appear on roughly half of all searches. Individual publishers have reported organic Google traffic falling by half or more year over year.
No. Ranking better cannot beat an AI answer that resolves the query above your listing — you can hold the top position and still lose the click, and Reddit proves that even maximal search favor and AI citation do not restore it. It is a distribution problem, not a ranking one. The fix is to stop depending on any borrowed referral channel and to move distribution onto surfaces you occupy directly and audiences you own, so no single intermediary's product decision can reprice your reach to zero.
Diversify off borrowed referrals on three fronts. Publish natively across the platforms where attention already lives, so discovery happens in-feed rather than via a Google click — including participating directly on communities like Reddit as a first-party presence, not just hoping search forwards visitors from it. Own an audience no gatekeeper controls, primarily email. And build a recognizable brand and persona so you are named and cited directly by both people and AI systems. The obstacle is production volume, which is where a generation-and-publishing engine changes the math.
Google's AI Overviews answer queries on the results page, so the same ranking sends far fewer clicks — and the loss reaches even Reddit, the platform Google most elevated, licensed for AI training in a reported $60-million-a-year deal, and cites most inside AI answers. In 2026 Reddit's stock still fell hard after its CEO warned AI Overviews made search referrals volatile. The lesson of SEO versus AI Overviews is not "rank better" but that any distribution built on borrowed referrals is fragile; own your presence across surfaces and an audience no gatekeeper controls.
Get started → · ← All guides · Compare Kompozy vs other tools