Google now treats AI citation gaming as link spam. Anyone actually working AI Mode/Overviews yet?

Caught this in the June updates and it's been rattling around my head. Google confirmed its spam policies now cover AI search -> buying or manipulating citations in AI Overviews and AI Mode gets treated the same as inauthentic backlinks. Same framework, same risk. They also changed how links and citations show up inside those answers.

So the citation game in AI answers is officially the link game now, same rules attached.

What I'm curious about from anyone actually in it: are you doing anything deliberate to show up in AI Overviews or AI Mode yet? And does it feel like the same work as ranking a page, or a different animal entirely?

Zero-click is sitting around 68% this year, so "rank it and wait for the clicks" clearly isn't the whole story anymore. Genuinely want to hear what's working before I sink more testing time into it.

yeah, i'd treat it as the same foundation but a different selection layer. getting mentioned by pages google already trusts for the query seems safer than trying to manufacture citations directly.

I'm not yet doing much, already made plans on how to go about it for one of my projects, but I think people would still continue to do what they want to. AI citations are recommendations is unlike search where Google held/holds an absolute monopoly. Many people won't mind if they'll consistently get cited by other AIs and drive traffic.

From my experience buying listing guest posts and creating links on a websites that Google rank in its AIO - is a good option.

I’m starting to look at it like ranking still the base game fr, AI just threw a whole extra layer on top of it. Feel like Google ain’t just asking, “can this page rank?” no more, they also asking, “would I actually trust pulling this into an answer?”

Lowkey I think folks finna waste a lotta time tryna brute-force citations the same way they was out here spamming links back then. Me personally, I’d rather build stuff that naturally get referenced case studies, stats, calculators, original data, actual useful pages and all that. If AI keep pulling from trusted sources, then being quote-worthy probably matter more than stuffing authority signals everywhere and hoping it stick.

Still testing tho. Anybody seeing pages that barely ranking but somehow still getting pulled into AI Overviews consistently? That part got me curious.

feels more like building authority than chasing rankings, at least from what i've seen so far

Appreciate everyone jumping in, let me actually go through these because you all hit different parts of it.

rohit755755 said:

getting mentioned by pages google already trusts for the query seems safer

Click to expand...

This is basically the whole thing in one line. "Different selection layer" is exactly it, the crawl and rank work gets you eligible, but getting picked is a separate decision on top. And earning the mention on a page Google already trusts is the durable version. Manufacturing it is the fragile one.

HenryObi said:

AI citations are recommendations is unlike search

Click to expand...

The no-monopoly point is underrated. It isn't one SERP anymore, it's ChatGPT, Gemini, Perplexity and Overviews all picking independently. That cuts both ways though: more surfaces to get pulled into, but also more places a gamed citation gets discounted the second any one of them tightens the screws.

uniteddare said:

buying listing guest posts and creating links on a websites that Google rank in its AIO

Click to expand...

This works right up until it doesn't. A mention on a page Google already pulls into AIO can ride you into the answer short term, no argument. But that's the exact pattern they just folded into the link spam policy, you're manufacturing the citation instead of being the thing worth citing. When they retune it, and they will, the bought mentions are first to get discounted.

qahnaxy said:

pages that barely ranking but somehow still getting pulled into AI Overviews consistently

Click to expand...

Yeah, and it's the most telling part. The pages I see getting pulled aren't the highest-DR ones, they're the ones with a clean extractable answer sitting right there: a table, a straight definition, a number in context. I've got a low-DR e-juice store (DR 21, ranks nowhere near page one for half its terms) whose comparison pages still get quoted into overviews well above their weight, purely because the answer is a clean lift. Authority gets you crawled. Being quotable gets you cited.

seo mastermind said:

still basically SEO, just showing up in a new surface

Click to expand...

Half with you. The plumbing is 100% SEO, index it, make it trustworthy, none of that changed. But "new surface" undersells the selection bit: a page can rank fine and still never get quoted because there's no clean line to lift out of it. That part is genuinely new, and it's where I'm spending my testing time.

Crypto Man said:

feels more like building authority than chasing rankings

Click to expand...

Yeah, with one tweak: it's less raw authority and more trusted AND quotable. Plenty of high-authority pages never get pulled because they bury the answer in 1,500 words of throat-clearing. Authority gets you considered. Being liftable gets you used.

Net for me: I've stopped treating position as the only scoreboard. With zero-click sitting where it is, "how often are we the one getting quoted" is the number I actually watch now, and so far it tracks usefulness a lot more than it tracks authority.

Hetneo said:

The no-monopoly point is underrated. It isn't one SERP anymore, it's ChatGPT, Gemini, Perplexity and Overviews all picking independently. That cuts both ways though: more surfaces to get pulled into, but also more places a gamed citation gets discounted the second any one of them tightens the screws.

Click to expand...

Exactly!

Many people won't mind being penalized by Google if they can get equivalent or even better traffic from ChatGPT and others. Monopoly is dying out as far as AI is concerned.

Hetneo said:

This works right up until it doesn't. A mention on a page Google already pulls into AIO can ride you into the answer short term, no argument. But that's the exact pattern they just folded into the link spam policy, you're manufacturing the citation instead of being the thing worth citing. When they retune it, and they will, the bought mentions are first to get discounted.

Click to expand...

Well, it depends. If the link has topical relevance, I doubt it can hurt your website. The logic behind this is the same as we've built our links previously - you need to find relevant website or page and create relevant link. Thats the best you can get.

HenryObi said:

Many people won't mind being penalized by Google if they can get equivalent or even better traffic from ChatGPT and others.

Click to expand...

The hedge is real, the sizes aren't. Conductor put numbers on this across 13,770 domains and 3.3 billion sessions: AI referrals came out at 1.08% of all traffic, and ChatGPT is 87% of that slice. Organic search still sits at 30-40% of traffic for most industries. So "equivalent or better traffic from ChatGPT" is a 2028 sentence, not a 2026 one. Growth is real, ChatGPT referrals roughly tripled year over year, but that's tripling from tiny. Eat a Google penalty today and you're swapping a 40% channel for a 1% channel.

Where I'm fully with you: nobody optimizes for one referee anymore. Which is exactly why I'd rather be the source these things pull from than rent mentions on pages that get pulled. Same asset works on every surface, and no single referee can take it away.

uniteddare said:

If the link has topical relevance, I doubt it can hurt your website.

Click to expand...

Relevance is necessary, it just isn't immunity. PBNs already taught everyone this lesson: perfectly relevant links on perfectly relevant sites still burned when the network got flagged, because Google wasn't judging the individual link, it was judging how the links came to exist. Same deal here -> the policy language targets manufactured citations, not off-topic ones. Relevance decides how good the camouflage is. The buying pattern is what they're hunting.

Where you're right: if someone's doing it anyway, relevant pages with real traffic that already show up in answers is the strongest version of the play. I just wouldn't build next quarter on links that only exist because money moved, on the one surface Google just told everyone it's policing.

@oanhka393 fair enough, most people will keep running whatever they were running. The update doesn't stop anyone, it just reprices getting caught.

Have anyone tried Google Preferred Source? is it helping AI overview?

Hetneo said:

Caught this in the June updates and it's been rattling around my head. Google confirmed its spam policies now cover AI search -> buying or manipulating citations in AI Overviews and AI Mode gets treated the same as inauthentic backlinks. Same framework, same risk. They also changed how links and citations show up inside those answers.

So the citation game in AI answers is officially the link game now, same rules attached.

What I'm curious about from anyone actually in it: are you doing anything deliberate to show up in AI Overviews or AI Mode yet? And does it feel like the same work as ranking a page, or a different animal entirely?

Zero-click is sitting around 68% this year, so "rank it and wait for the clicks" clearly isn't the whole story anymore. Genuinely want to hear what's working before I sink more testing time into it.

Click to expand...

There is no separate GEO playbook that's worth chasing yet.

What has been working for me is keeping things simple: structure your content so it answers questions clearly, support it with a handful of relevant citations, and focus on building genuine authority. Trying to manipulate AI citations directly isn't a sustainable strategy. Instead, invest in high-quality backlinks, publish useful content, earn brand mentions on authority websites, and include appropriate citations where they add value.

That's the approach I've been using, and it's consistently delivered better results than chasing AI-specific hacks.

google pulls relevant content in AIO / ai mode.
site owners can manipulate that using some tricks.

I've found that the fundamentals still matter the most. Instead of chasing AI citations, I'm focusing on creating genuinely helpful content, covering topics in depth and building authority through trusted mentions and real expertise. AI Overviews seem to reward the same signals that make a page valuable to users, so I'm treating them as an evolution of SEO rather than a completely different game.

I've started treating AI visibility as an extension of SEO rather than a completely new strategy. Instead of chasing citations directly, I'm focusing on publishing genuinely useful, original content that's easy to understand and backed by real expertise. From what I've seen so far, the same trust signals that help pages rank also seem to increase the chances of being cited in AI answers, even if the end goal is a little different.

Hetneo said:

Caught this in the June updates and it's been rattling around my head. Google confirmed its spam policies now cover AI search -> buying or manipulating citations in AI Overviews and AI Mode gets treated the same as inauthentic backlinks. Same framework, same risk. They also changed how links and citations show up inside those answers.

So the citation game in AI answers is officially the link game now, same rules attached.

What I'm curious about from anyone actually in it: are you doing anything deliberate to show up in AI Overviews or AI Mode yet? And does it feel like the same work as ranking a page, or a different animal entirely?

Zero-click is sitting around 68% this year, so "rank it and wait for the clicks" clearly isn't the whole story anymore. Genuinely want to hear what's working before I sink more testing time into it.

Click to expand...

In my opinion the fundamental change is not in asking how to rank but asking how to be a source worth being quoted. Ranking are important, no doubt about it hoever it is data, information nd expertise that will make AI visible.

It feels like Google is rewarding quality more than ever Chasing AI citations doesn't seem like a sustainable strategy I'm focusing more on useful content and natural visibility now

Been off this thread a few days and good stuff piled up. @benclud, @ElaraFrost and @seopowergirl all landed on the same spot: it's an evolution of SEO, not a new religion. Agreed. My only gripe with "just make quality content" is that it's true and also useless as an instruction, so let me try to say the part that actually changed.

Linkzo said:

There is no separate GEO playbook that's worth chasing yet.

Click to expand...

Respect for the "no separate playbook" line, that's the bit nobody flogging a GEO course will say out loud. Only thing I'd add: everything you listed, the backlinks, the useful content, the brand mentions, gets you into the running. It doesn't decide who gets pulled into the answer. I keep seeing DR 60 pages that never get quoted because the answer's buried under 1,500 words of throat-clearing, and scrappy little pages getting lifted constantly because the sentence is just sitting there clean, ready to copy. That selection step on top is the genuinely new bit.

Reezaniu said:

the fundamental change is not in asking how to rank but asking how to be a source worth being quoted

Click to expand...

Yeah. And the good news is you can measure it, which you can't really do with "authority". I stopped treating rank position as the only scoreboard and started tracking how often we're the one getting quoted. That number moves when I make a page more liftable and it barely tracks DR at all. Rankings are still the plumbing, I'm not throwing them out, but quotable is the lever I can actually pull.

GhostlyGardener said:

just EEAT + entity SEO + citation-worthy content applied to a new SERP feature

Click to expand...

This is the one to read twice. The thin best-of pages getting torched is the lesson half the forum is about to learn the expensive way, that's the exact footprint they folded into the spam policy. One tweak though: schema and clean Q&A blocks help Google parse you and consider you, sure, but what ends up in the answer is the plain sentence in your copy, not the markup around it. I'd write the liftable sentence first and bolt the schema on after, not the other way round.

Fionacc said:

Have anyone tried Google Preferred Source? is it helping AI overview?

Click to expand...

Careful, that's a different surface. Preferred Sources is the thing where a user ticks a box to see more of a publisher they already like in Top Stories. So it helps you reach people who already picked you, which is a branded-demand play, handy but narrow. I've seen nothing that says it feeds who AI Overviews decides to cite. If AIO is the target, the liftable-answer stuff above does a lot more than getting ticked as someone's preferred source.

Net for me, same as when I started the thread: I'd rather be the thing these engines quote than rent a mention on something they already quote. One asset, works on every surface, and no single one of them can switch it off. That's where my testing time is going, not into citation placements.

I've moved from "backlinks" to "citations + branding + backlinks" as it's becoming the new normal. AI tools are as valuable as search engines for long-term businesses. It's too early for results but I'm sure that things will eventually fall in line.

too early for results

Click to expand...

The stack is right but "too early for results" is only half true. The citation leg is measurable today: take the ten questions your money pages answer, ask them in the AI surfaces your buyers actually use, note who gets cited. Do it weekly and you have a trendline long before anything shows in analytics. What we keep seeing is that the sources LLMs lean on are often not the pages winning the classic SERP for the same query, and that gap is exactly where the opportunity hides.

Curious what your branding leg looks like in practice. Actual PR, or entity hygiene (consistent mentions, sameAs, getting the knowledge graph to agree on who you are)?