Two days ago, a senior SEO practitioner with 115,000 followers published a Substack article that should've been a wake-up call for every founder chasing AI search visibility. The thesis was simple: the GEO tactics everyone's racing to implement are actively destroying their Google rankings. And the kicker? Killing your SEO kills your AI visibility too.
The timing wasn't random. The SEO community had spent the past week watching case study after case study of sites tanking their organic traffic while chasing "AI-optimized" content strategies. The conversation blew up. Tens of thousands of views. Hundreds of industry veterans jumped in.
I've been tracking both sides of this for months through our AI visibility monitoring, and the pattern is unmistakable. Let me walk you through what's actually happening.
The GEO Gold Rush (and Why It's Getting Reckless)
Here's what I keep seeing. A SaaS founder reads that ChatGPT referral visitors convert 4.4x better than organic Google visitors. They see the headlines about AI search eating Google's lunch. So they pivot hard. They hire a "GEO agency" or follow a viral thread about optimizing for Perplexity, and they start implementing every tactic they find.
Three months later, their Google organic traffic is down 30%. Their AI citations haven't moved. They're worse off on both channels. SEO veterans call this the "Mount AI" pattern: rapid AI-content growth followed by an equally rapid crash, with years of organic authority wiped out in weeks.
GEO itself isn't bad. But most GEO advice being sold right now treats Google traffic like it's already dead. It's not. For the vast majority of B2B SaaS companies, Google still drives 50-90% of organic leads. AI search referrals, for all the hype, account for less than 1% of traffic to most websites.
That's not an opinion. That's what we see across every site we track.
Here's the other thing that keeps coming up. Most GEO "case studies" fall into one of two traps. The first is relabeling SEO tactics that have worked for years (structured data, FAQ schema, clear headings, direct answers) as "novel GEO innovations." The second is claiming GEO credit for AI citations that were almost certainly driven by organic rankings the site already had before any GEO work started. A brand with years of authority and strong backlinks starts appearing in ChatGPT responses and decides their new GEO agency made it happen. More likely, their existing SEO visibility is what got them into the indexes feeding those AI products in the first place.
Four GEO Tactics That Actually Tank Your Rankings
Not all GEO tactics are dangerous. Some overlap perfectly with good SEO. But these four? These are the ones causing real damage.
1. Mass-producing thin "AI-optimized" pages
The most common play: spin up dozens of pages designed to answer specific AI queries. Each page is 400-600 words, stuffed with direct-answer formatting, and covers a narrow question.
Google calls this "scaled content abuse." Doesn't matter if a human wrote it or GPT did. If you're publishing 50 thin pages in a month that add no unique insight, Google's quality systems will catch it. We've watched sites lose 40-60% of their indexed pages within two core updates after implementing this exact strategy.
The GEO logic sounds right: "AI models want direct answers, so give them direct answers on dedicated pages." But Google wants depth, uniqueness, and genuine expertise. A page that exists only to be cited by ChatGPT is a page Google has every reason to demote.
Know exactly what AI says about your competitors.
RankControl's Recon Agent monitors competitor citations across ChatGPT, Perplexity, Claude, and Gemini. See where they show up and you don't.
2. FAQ schema carpet-bombing
FAQ schema is fantastic for AI citations when used properly. The problem starts when someone reads "AI models love FAQ schema" and decides to slap 15 FAQ blocks on every page, including pages where the questions are barely relevant.
Google's been cracking down on FAQ rich results since 2023. Overusing FAQ schema doesn't just waste your structured data budget. It can flag your site as manipulative. And here's the part nobody mentions: if Google stops trusting your structured data, AI crawlers that rely on Google's index to discover your content will see you less, not more.
Use FAQ schema on pages where the questions are genuinely useful. Four to five questions per page. Skip it on product pages, landing pages, and pages where the FAQs are clearly just SEO padding.
3. Rewriting existing high-ranking content for "AI readability"
This one hurts the most because the intention is good. You have a page ranking #3 for a competitive keyword. You read that AI models prefer "definition-first content" with "direct answers in the opening sentence." So you rewrite your intro, strip out the nuance, and turn your expert analysis into a Wikipedia-style summary.
Your AI citation rate stays flat. Your Google ranking drops to page 2.
I should probably mention why this happens. Google's ranking systems reward content that demonstrates E-E-A-T: experience, expertise, authoritativeness, and trustworthiness. That detailed, opinionated, experience-driven content is exactly what makes a page rank well. When you strip it out to make the page "AI-friendly," you're removing the signals Google uses to rank you.
The fix is additive, not subtractive. Keep your expert content. Add a direct-answer summary at the top. Add an FAQ block at the bottom. Don't gut what's already working.
4. Excessive alternative/comparison page spam
This tactic exploded in the last six months. The playbook: create dozens of "[Your Product] vs [Competitor]" and "[Competitor] alternatives" pages, each targeting a different AI query pattern.
A few of these pages? Great. Standard competitive SEO that also works for AI citations. Twenty or thirty of them, all following the same template, with thin content that barely differentiates between pages? That's a quality problem. We've seen sites publish 40+ comparison pages in a single quarter and then wonder why their domain authority dropped like a rock.
The GEO case studies promoting this approach typically fall into one of two traps. The first is repackaging the same content with different headers. The second is creating comparison pages for competitors your audience doesn't actually compare you against. Both patterns signal low-quality programmatic content to Google.
15+ content types. Published on your domain. Matched to your brand.
Guides, comparisons, listicles, case studies, and more. RankControl generates content that gets cited by ChatGPT, Claude, Perplexity, and more.
The Irony Nobody's Talking About
Backing up a step. There's a fundamental misunderstanding driving this entire problem.
Most people assume Google and AI search are separate channels with separate rules. Optimize for one, accept tradeoffs on the other. That framing is wrong.
71.7% of ChatGPT citations come from pages with organic search presence (we wrote about this in our 48-hour citation experiment). Perplexity and Gemini show similar patterns. AI search engines don't maintain their own independent index of the entire web. They lean heavily on pages that already demonstrate authority through traditional search signals.
So when you tank your Google rankings with aggressive GEO tactics, you're not trading Google traffic for AI traffic. You're losing both.
This is the part that makes me genuinely frustrated with the GEO hype cycle. The case studies showing "We 10x'd our AI citations!" almost never show what happened to Google traffic during the same period. When they do, the numbers often tell a different story.
There's another layer to this that most GEO vendors won't tell you. Ahrefs research shows that 45.5% of AI citations get replaced when you regenerate the same query. The same brand appearing in an AI response one hour might disappear the next. So even when GEO tactics "work," the visibility you gain is fundamentally unstable compared to a Google ranking that holds for months.
Our citation tracking shows a consistent pattern: sites that maintain strong Google rankings and add AI-friendly content structure on top see steady AI citation growth. Sites that sacrifice Google performance to chase AI citations see initial spikes followed by declines in both channels.
And the timing pressure is real. ChatGPT's quality filters are visibly tightening. We're tracking sites that saw massive declines in ChatGPT mentions and citations over the past month, with ChatGPT surfacing much more relevant content than before. Meanwhile, Bing recently updated its webmaster guidelines with explicit language about AI search spam and manipulation, and stated plainly that SEO and GEO work together. Google is even developing opt-out controls to let publishers exclude themselves from generative AI features in Search. The platforms are telling you where this is heading.
What Actually Works: The Both-And Approach
The good news? The overlap between good SEO and good GEO is about 70-80%. Most of what helps you rank on Google also helps you get cited by AI. The last 20% is where you need to be careful.
Tactics that help both channels:
- Clear, direct answers at the top of each major section (inverted pyramid)
- FAQ schema on pages with genuinely useful questions (4-5 per page)
- Strong topical authority through content clusters
- Fast page speed and clean technical SEO
- Properly configured
robots.txtthat doesn't block AI crawlers llms.txtfile pointing AI models to your best content- Article schema with accurate
dateModifiedtimestamps - Comparison tables and structured data where natural
Tactics that help GEO but risk SEO:
- Mass-producing thin answer pages (safe at 3-5 per month, dangerous at 20+)
- Rewriting existing content to be more "AI-readable" (additive is safe, subtractive is risky)
- Over-indexing on FAQ schema across your entire site
- Creating dozens of templated alternative/comparison pages
- Removing opinion, nuance, and experience from content to make it "cleaner" for AI
The difference between the two lists comes down to one question: does this tactic add value for human readers, or does it only exist to game an AI model?
If you'd be embarrassed to show the page to a customer, it probably shouldn't exist.
The Time Cost of Getting This Wrong
For what it's worth, let me be real about what recovery looks like.
If aggressive GEO tactics trigger a Google quality demotion, you're looking at 3-6 months of recovery time. That's not a guess. That's the typical timeline we see after a core update hits a site that over-indexed on thin AI-optimized content. During those 3-6 months, you're losing Google traffic and AI citations simultaneously.
The math is brutal. Say your site gets 10,000 organic visits per month with a 2% conversion rate. That's 200 leads per month. A 30% traffic drop costs you 60 leads per month. Over a 4-month recovery window, that's 240 lost leads.
Meanwhile, the AI search traffic you were chasing? At less than 1% of total traffic for most sites, even a perfect GEO strategy might add 100 visits per month. At 4.4x the conversion rate, that's about 9 extra leads per month.
You do the math. Nobody should be trading 60 leads per month for 9.
What to Do Instead (the Checklist)
Here's the approach that actually works without putting your Google traffic at risk.
Start with measurement. Before you change anything, baseline both channels. Track your Google rankings for target keywords and monitor your AI citations across ChatGPT, Perplexity, Claude, and Gemini. You can't optimize what you can't measure. And the hard part isn't doing this once. It's catching the moment something stops working.
Audit your content structure, not your content quality. Look at your top 20 pages by Google traffic. For each one, check: does the first paragraph contain a direct, quotable answer? Is there FAQ schema? Are there comparison tables where relevant? Most pages need structural additions, not rewrites. Our structured data schema blueprint has the specific JSON-LD templates that work for both Google and AI search.
Build topical authority through depth, not volume. Instead of 20 thin pages, publish 5 genuinely useful pages that each go deep on one question. Link them into a cluster. AI models evaluate whether you have depth on a topic, and Google evaluates the same thing. This is the overlap that works for both.
Monitor weekly, not monthly. AI search behavior changes fast. A page that gets cited this week might not get cited next week. Google rankings fluctuate around core updates. Weekly monitoring lets you catch problems before they compound. Monthly check-ins are too slow for this space.
Total time for this approach: 10-15 hours for the initial audit, then 3-5 hours weekly to maintain. Not trivial. But a lot cheaper than 3-6 months of recovery from a quality demotion.
You can do all of this manually. Or RankControl's agents can run the audit, track both channels, and alert you the moment something shifts, every week, while you focus on your product.
Your competitors are getting cited by AI. You're not.
Every day without citation tracking is a day your competitors pull ahead in ChatGPT, Perplexity, and Claude.
