EDITOR'S NOTE 👋
Hey there 👋
You rank #1, you own the featured snippet on Google, but then you ask ChatGPT the same question, and you're nowhere in the answer.
I get sent this problem a lot. Usually, it's from people who assume something's broken.
But nothing's broken. The two systems just work differently.
SE Ranking looked at 10,000 keywords and found that the exact URL ranking in Google's organic top 10 shows up in AI Mode's citations about 14% of the time, while a Moz study put the same figure at 12%.
Now, that doesn't mean rankings stopped mattering. You should still make sure you’re optimized to rank in traditional search, but some things you need to approach differently if you want to increase your brand’s AI search visibility.
In this week’s issue, we're going to break down why ranking #1 is no longer enough, what the "SEO-GEO gap" looks like, and the specific tactics you need to adopt to ensure your content gets cited by AI engines.
Let's go! 🚀
TL;DR 📝
Ranking #1 doesn't buy you the citation. About 12% of AI Mode citations are the exact URL ranking organically for that query, and roughly 22% share a domain with it.
AI Mode breaks your question into several background searches and answers from all of them, so pages that never rank for the head term can still end up cited.
Building a separate page for every fan-out variation is the one tactic Google has named as a spam risk. Cover the intents inside content that earns its place.
Original data earns citations at very different rates to generic education. Across 10 sites, trends and analysis posts hit a 78% citation rate against 12% for how-to content.
Reddit, YouTube and LinkedIn show up heavily in AI citations, so your off-site presence now shapes what the answer says about you.
NEWS YOU CAN USE 📰

Only 12% of AI Mode citations match the organic SERP. Moz found the overwhelming majority of AI Mode citations aren't the URL ranking organically for that exact query. SE Ranking's separate analysis of 10,000 keywords landed close, at 14% exact URL overlap with the top 10 and 21.9% at the domain level. Traditional rankings still matter, but they're no longer the only route in. [Source: Moz]
AI Overviews now show on 43% of US searches. Up from 15% a year earlier, according to Similarweb's 2026 Generative AI Landscape report. Worth keeping the two surfaces straight: AI Overviews are summaries inside conventional results, while AI Mode is the conversational interface for longer questions. Readers can jump from one into the other without losing the thread of their original search. [Source: AI News]
Shopify says AI search is adding demand rather than replacing it. On the Q2 earnings call, president Harley Finkelstein said AI has become "a complement to search, rather than a substitute for it." AI-driven traffic and orders tripled year over year, and 75% of AI-attributed purchases landed outside the top 100 product categories. If you sell anything with a long tail, that last number is the one to sit with. [Source: TechCrunch]
WHY THE SEO-GEO GAP EXISTS, AND HOW TO CLOSE IT 🧠
Ranking #1 means Google's systems judged your page highly relevant and useful for that query. That judgment still feeds AI Mode, which draws on the same index and the same quality systems.
What changes is what happens next. AI Mode doesn't pick a winner. Instead, it assembles an answer, often from searches you never made, and the citations that come out the other end can look very different from the results page you're winning.
Four reasons that gap opens, and what I'd do about each.
Query fan-out
Type a question into AI Mode, and it doesn't just run your question. It decomposes it and runs related searches in the background.
Ask "what's the best CRM for a 50-person remote marketing agency?" and the engine works through something more like:
best CRM for marketing agencies
CRM software remote teams 50 employees
HubSpot vs Salesforce mid-market agencies
reddit best agency CRM 2026
Then it pulls passages from those and synthesizes one answer. You might get cited off the head term. If a competitor owns the comparison queries and the forum threads, they show up more often across the fan-out, and the answer leans their way.
What I'd do, and what I wouldn't. The advice going around is to spin up a page for every fan-out variation. Don't. Google has said directly that building content around every possible variation of how people might search, mainly to influence AI results, can violate its scaled content abuse policy. That's a named spam risk from the company running the engine.
Cover the adjacent questions and comparison intents where they genuinely help someone. Sometimes that's a section inside an existing guide. Sometimes it's a page of its own, when it serves a real need on its own. The test is whether it would still deserve to exist if AI search vanished tomorrow.
Retrieval works on passages
AI systems retrieve and cite specific sections rather than whole pages. That's real, and it's why a page can rank beautifully while the paragraph that actually answers the question sits at number 14 and never gets pulled.
I'd temper this, though, because it's the area with the most bad advice attached. Conventional Google Search already understands passages too, so this isn't a brand new dimension. And Google has said there's no requirement to chop your content into small pieces for AI to understand it.
What I'd do: Answer the page's core question early and plainly, and write paragraphs that still make sense if someone lifts them out of the page. Clear headings help. That's good writing, and it happens to make retrieval easier.
Treat the "answer in the first 100 words" idea as something to test on a few pages rather than a rule. I've seen it help. I can't show you it caused the help, and anyone who tells you they can is selling something.
Original data beats good explanations
Generic educational content is the bread and butter of traditional SEO, and it's what models least need you for. Nothing has to cite you to explain what B2B marketing is.
What they can't generate is your original data. Adam Gnuse analyzed 10 sites and 150,000 pages, and the spread is stark. Trends and analysis posts earned an LLM citation rate of 78%. Data-based year-in-review posts hit 61%. Educational how-to content managed 12%.
Tools are more interesting than they first look. Interactive tools had the highest per-page rate of picking up LLM traffic in that dataset, but their normalized volume was the lowest of the page types measured, at 9.8 LLM sessions per 1,000 organic sessions against 23.4 for articles and 29.4 for service and product pages. So a tool is likelier to get picked up, and gets less volume when it does.
What I'd do: Move budget out of top-of-funnel glossaries and into original research and proprietary data. If you've built a calculator or an assessment, give it a real name people can ask for, because an engine can only recommend a tool it can name.
Off-site is now part of your search performance
AI engines lean on places where people talk to each other, though how much varies by engine and by query set.
Peec AI's aggregate analysis of 30 million citations put Reddit first, YouTube second, and LinkedIn third. Inside AI Mode specifically, the mix shifts, with SE Ranking's June 2025 sample showing Indeed, Wikipedia, Reddit and YouTube among the most-cited external domains. Worth knowing that in that sample the top 10 domains together accounted for 11.9% of citations, so this is a meaningful slice rather than the whole picture.
The practical version: a highly ranked page of yours can end up in an answer next to forum or video commentary that says something different, and the model weighs both.
What I'd do: Treat off-site visibility as a search deliverable rather than a PR nice-to-have. Answer questions in the threads where your category actually gets discussed, publish the substantive version on LinkedIn, and put your best explanations on YouTube.
THIS WEEK'S PROMPT 🤖

Use this on a page that ranks well organically and gets cited nowhere.
The Scenario: You've got a post sitting at #1 for a keyword you care about, and it isn't showing up when you run the query through ChatGPT or AI Mode.
The Prompt:
You are a Generative Engine Optimization specialist. I'm giving you the text of a blog post that ranks #1 organically but earns no citations in AI search. Rewrite it so the answer is easier to find and quote.
Context:
The post is a comprehensive guide to [Topic]
It ranks #1 organically and gets no AI citations
[PASTE YOUR ARTICLE TEXT HERE]
Tasks:
Identify the core question this page answers. Draft a 50 to 75 word answer to it, plain and direct, for placement near the top.
Rewrite the H2s and H3s as the natural-language questions someone would actually type or say.
Find three paragraphs that hold real information but only make sense in context. Rewrite each so it stands on its own.
Flag anywhere prose would work better as a table or a list.
On measuring it. Track citations and referral traffic as two separate things, because plenty of citations never get clicked. Run the same prompt several times across different days and clean sessions before you change anything, and again afterward. AI answers are volatile: SE Ranking found only 9.2% exact-URL overlap between three AI Mode runs on the same day. A two-week comparison is an early signal, not proof the rewrite did it.
TOOLS WE USE ⚒️
These are the most popular AI tools we use at Rise Up Media. If you're not using them already, they're worth a look.
LLMRefs: We've recently started using LLMRefs to track our clients' AI Search visibility.
Manus AI: General-purpose AI agent we love (and use to create this newsletter)
n8n: Open-source automation (if you like that sort of thing)
OpusClip: Auto-clips long videos into shorts (and is really good at it)
Buffer: Manage all your socials (with a sprinkle of AI) in one place.
Full disclosure: some links above are affiliate links. If you sign up, we'll earn a small commission at no extra cost to you.
WRAPPING UP 🌯
Nobody wants to restructure a page that's already winning. I understand the reluctance, and I'd feel it too.
But the thing evaluating your content changed, and it's asking a different question than Google's ranking systems ask. Google wants to know which page is the best result, while an AI engine wants to know which paragraph answers this specific piece of a question it broke apart three seconds ago.
Your rankings still feed that, but they're just no longer the whole of it.
Keep aiming for that #1 ranking but start writing for the fan-out as well.
Until next time, keep exploring the horizon. 🌅
Alex Lielacher
P.S. If you want your brand to show up in Google AI Mode, ChatGPT, and Perplexity, reach out to my agency, Rise Up Media. That's what we do!


