Your rankings are holding. Your clicks aren’t.
Google’s AI features answer the inquiry on the results page, and the searcher never leaves. From January to April 2026, 68% of U.S. Google searches ended without a click. And small of Google’s traffic reaches the open web than it used to.
Because for a expanding portion of B2B buyers, the hunt doesn’t commencement on Google at all. It starts in an AI chatbot.
Google’s results page, its AI features, and the standalone LLMs all drag from a mix of sources. Reddit threads. YouTube videos. Review profiles. LinkedIn posts. Comparison pages.
That’s your integrated growth stack. It’s additionally the mix your buyers use to build a shortlist.
If you lone optimize your site, you’re competing for a shrinking piece of the pool. To remain competitive, SEO has to develop beyond on-site.
Today, I’ll display you which sources all exterior leans on and what to build first.
Search Is Splintering Across Surfaces
For our State of B2B AI Discovery 2026 report alongside AirOps, we tracked approximately 380,000 AI answers covering 81 endeavor application topics throughout six AI surfaces. Unless connected otherwise, the figures below arrive from this study.
Let’s commencement alongside what didn’t change. Product pages motionless get cited additional than item else.
Across all AI exterior we tracked, vendor sites, integrators, and tool directories took the largest portion of citations. That ranged from 52% on Perplexity to 72% on Gemini. Review sites added another 9% to 21%.
Remember: Vendor portion isn’t yours by default. It includes your competitors’ sites and the application directories that catalog you side-by-side.
This is the stack’s foundation, and it’s crowded. Your merchandise pages vie alongside your rivals’ pages. The difference pages written concerning you vie alongside the ones you compose yourself.
The layers complete the basis are anywhere hunt starts to splinter.
Google’s AI Overviews and AI Mode dirt their answers in live Google Search results. They inherit what already ranks. ChatGPT, Claude, Gemini, and Perplexity run their own retrieval.
The citations divided correct alongside that line.
Google’s two AI surfaces put 7% of their citations into social platforms akin YouTube and LinkedIn. ChatGPT, Gemini, and Claude sit at 0.3% or below.
Domain by domain, the gap gets sharper:
- Reddit appears in 52% of AI Mode’s citing answers and 39% of AI Overviews. ChatGPT cites it in 21%, and Claude didn’t citation it at all.
- YouTube appears in 47% of AI Mode answers and 38% of AI Overviews. ChatGPT and Claude: basically zero.
- LinkedIn reaches 13% of AI Mode answers, 9% of AI Overviews, and 6% of Perplexity’s.
- G2 lands between 23% and 30% on five of the six surfaces. It’s the one domain all exterior agrees on.
So “AI visibility” is at smallest two games.
A brand can be everyplace in Google’s AI answers but invisible in ChatGPT and the another standalone models. Or the reverse. Using a blended visibility score hides all one’s contribution.
See also: Wikipedia, Podcasts & Reddit Now Drive AI Citations, Ask Your PR Team How They Did It
Build Your Organic Growth Stack Layer By Layer
Organic visibility now comes downward to exterior coverage: showing up on the domains all AI exterior trusts most.
Your new integrated growth stack has to location the two internets. Here’s how to do it.
See also: What Search Engines Trust Now: Authority, Freshness & First-Party Signals
1. Start With The Pages You Own
AI tools are catching on accelerated alongside B2B buyers. According to G2, 51% of B2B application buyers now commencement their investigation in an AI chatbot additional frequently than Google, up from 29% a twelvemonth earlier.
But Google motionless dominates hunt overall. It handles approximately three-quarters of all desktop searches. AI Overviews solitary attain 2.5 billion monthly users and another billion use AI Mode.
I motionless think ranking matters. It fair doesn’t justify a citation anymore. AI Overviews have been pulling small citations from top-ranking pages.
The standalone engines lean on merchandise pages too.
Gemini sends 72% of its citations to vendor and merchandise domains, during G2 shows up in fair 14% of its answers. Gemini is choosing between vendor pages, and yours are up against everyone else’s in your category.
Then there’s difference content.
Eighty-four of the 100 most-cited URLs in our data were either difference pages (59) or listicles (25). Together, they drew complete 66,000 citations. Whoever writes those pages shapes how AI frames your category.
A lot of teams spill prosperity into the blog and let merchandise and difference pages go stale. I’d flip that. Fix the pages AI already cites before you compose another post.
As James Scherer, Foundation’s VP of Strategy, puts it: “Editorial gets the budget; merchandise pages get the citations.”
Action: Identify the 10 most-cited difference pages and listicles in your category. Check how you’re framed in all one. Then publish your own difference leaf against your top three competitors.
2. Earn A Place In The Communities Your Buyers Frequent
Reddit was the most-cited external domain in our dataset, earning 159,000 citations throughout 12,500 distinctive URLs.
Reddit’s logged-out users, who mostly attain from hunt engines, grew 27% twelvemonth complete twelvemonth in Q2 2026. They now outnumber logged-in users, 77.7 myriad to 52.6 million. Threads position on Google. Google sends readers into the threads. Then Google’s AI features citation them.
The layers nourish all other.
But those layers can additionally change fast. After reports that ChatGPT was citing Reddit less, we extended our tracking and established its Reddit citation charge cut from 21% in July to 10% in August, during Google’s surfaces kept citing Reddit.
How much community platforms akin Reddit matter additionally depends on your category. Community citations ranged from 5% of the total for SEO platforms to 14% for EDR/XDR safety tools.
If you’re new to the platform, commencement alongside this SEJ AMA alongside a Reddit promotion veteran.
My biggest advice: Don’t treat Reddit akin a initiate channel. I’d commerce one viral thread for steady, helpful answers throughout 50.
Action: Spread your attempt throughout many threads and subreddits. Don’t bet on one. Track community citations weekly, since monthly reporting would have missed a autumn akin ChatGPT’s.
3. Show Up Where Categories Are Defined
LinkedIn is a Google-surface play. It reaches 13% of AI Mode answers, 9% of AI Overviews, and 6% of Perplexity’s. ChatGPT, Gemini, and Claude didn’t citation it at all.
Across B2B application categories, social is the second-largest citation origin following brand and merchandise content, at 6%.
LinkedIn is anywhere executives and practitioners set the category narrative. Google’s AI features choice up that narrative. It’s why employee amplification belongs in the stack.
Don’t autumn into the likes and follows trap. If the goal is getting cited, observe whether Google’s AI answers choice up the way your group depict the category.
Action: Pick the category questions you desire to own. Post on them all week from executives and practitioners. Measure the effect in AI Overviews and AI Mode, since that’s anywhere LinkedIn gets cited.
4. Make Video Content For Google’s AI Search Features
YouTube shows up in 47% of AI Mode’s citing answers and 38% of AI Overviews.
It’s tempting to peruse that as Google favoring its own property, but our data complicates that. Gemini is additionally a Google product, and yet YouTube showed up in 0.3% of its citations: approximately 17x small than on AI Overviews.
The difference is how all exterior finds its sources. AI Overviews and AI Mode are Search features. Google uses query fan-out, operating multiple connected searches to build all answer, so videos that already position in Search get pulled in.
A lot of B2B videos are made for group who already cognize your brand and product. But the videos that acquire a location in this tier answer the questions buyers ask before they’ve picked a vendor.
Action: Ship comparison, alternatives, and best-of videos. Title them the way your buyers expression the question.
5. Treat Review Profiles As Retrieval Infrastructure
The standalone engines lean difficult on assessment sites. Claude and Perplexity diagram concerning one in five citations from assessment domains.
Claude’s third-party citations are nearly entirely assessment platforms. G2 appears in 30% of the Claude answers that citation sources. Capterra appears in 13%, and Gartner and PeerSpot in concerning 9% each.
Category matters current too. Review citations ranged from 8% in managed databases to 25% in ATS and recruiting software.
Buyers rear this up. In G2’s 2026 research, assessment sites (38%) narrowly passed AI chatbots (37%) as the top origin shaping B2B application shortlists.
G2 isn’t a one-time push. For Claude and Perplexity, your overview is a live source, so run it akin one.
Action: Give your G2 overview a named owner and a monthly assessment target.
What To Build First
Nobody builds all tier at once. Run two checks before you start.
- Ask buyers which AI tools they use. Do it on revenue calls and during onboarding. If they use Google, prioritize the community, video, and social layers. If they use ChatGPT or Claude, prioritize your own pages and assessment profiles.
- Pull your category’s origin mix before you set a budget. Review citations sway 3.3x throughout categories. Community citations sway 2.5x.
Copy the incorrect integrated growth stack, and you end up investing in the incorrect channels. The AirOps data shows fair how much the origin mix changes from industry to industry, equal inside the SaaS space.
The managed repository niche is an explanatory example. A brand in this area that borrows an HR tech playbook pours prosperity into assessment sites, which acquire fair 8% of that category’s citations. Meanwhile, it starves communities, which acquire 13%.
Once you cognize which layers matter in your category, equivalent them to the squad you have. At all size, the inquiry is the same: which net are you motionless missing?
- If it’s fair you: Your own pages, affirmative the one tier your category leans on most. The basis tier gets cited on all AI surface, so it pays off whichever tool your buyer uses. A second tier you can’t keep up alongside fair goes stale.
- Three to five people: Add a second tier from the another internet. If you started alongside Reddit, which feeds Google’s AI answers, add assessment profiles, which nourish the standalone engines. A second tier on the identical flank lone makes you stronger anywhere you’re already visible. Give assessment profiles a named owner, since they go old fast.
- Eight or more: One owner per layer, affirmative one individual who reports citations by surface. At this size, the bigger hazard is drift. A blended mark hides it whenever one phase changes what it cites, the way ChatGPT did alongside Reddit.
Finally, measure by surface.
Run a fixed set of 100+ prompts on a schedule. Track citation charge and portion for all AI surface. Set an vigilant for any 20% week-over-week swing.
Watch opinion per platform, too. On the identical prompts, AirOps’ 100-point opinion scores ranged from 81 on Perplexity to 69 on Claude.
Then tie it rear to the numbers guidance already watches, akin tagged searches, referral traffic, and pipeline originated from your content.
Give it 12 months.
The archetypal fourth is slow. You’re fixing old difference pages and environment baselines nobody asked for.
By duration 12, it compounds. Your difference pages are the ones AI quotes in your category. Your assessment profiles remain caller without a scramble. Your threads and videos display up whenever Google answers your buyers’ questions. And whenever a phase changes what it cites, you capture it that week.
Last month, you built the engine. Now you cognize anywhere to item it.
Search gets you found. The stack gets you chosen.
More Resources:
- The 5-Pillar Framework For AI Content That Audiences Actually Trust
- So Your Traffic Tanked: What Smart CMOs Do Next
- The Same But Different: Evolving Your Strategy For AI-Driven Discovery
Featured Image: Roman Samborskyi/Shutterstock
