Mount AI Pattern: Rising AI Organic Traffic That Then Disappears
Growth hack? Write a ton of AI content, see your organic traffic rise, then watch it all disappear. That’s the Mount AI pattern- a phrase Glenn Gabe coined for the SEO problem that SaaS ecosystem is facing right now. And the evidence is piling up.
The anatomy:
- Steep climb in organic visibility over 3–6 months from mass-published AI content
- Traffic peaks as thousands of new pages rank broadly in Google’s initial indexing pass
- Equally steep traffic crash within 6–12 months as Google reassesses quality and originality
- Majority of sites end up below their pre-AI-content baseline
Lily Ray reviewed 220+ sites that launched major AI content initiatives. She discovered the following:
“Rapid growth in organic pages over six to twelve months; an organic traffic peak within roughly three to six months of the content peak; and then a steep decline in traffic that erases most of the gain (and frequently drops below the prior baseline) within the following year.”
The damage? 54% lost 30% or more of their peak organic traffic. 22% lost 75% or more.
Here’s what tripped up SaaS teams: AI content automation looks and feels like a growth lever. It’s fast. It’s cheap. The early results are stunning. But you’re not building defensible organic growth. You’re borrowing it from Google’s initial indexing pass. The real problem comes when Google re-assesses what you published.
Why This Crash Keeps Taking People by Surprise: Google Reassessment Issue
So why does the crash keep surprising people? Google doesn’t penalize you on day one for publishing AI content at scale—it indexes it broadly. Then it reassesses.
Essentially, it’s AI spitting out a bunch of content, that temporarily ranks high but then falls off because it doesn’t provide much information gain or originality.
Here’s what actually happens behind the scenes. Google’s initial indexation feels like a win. It proved true in an experiment by SE ranking, which found 2,000 AI articles got 71% indexed within the first month. However, after six months, only 3% were ranked in the top 100 . This isn’t a glitch – it’s Google’s re-assessor algorithm doing what it’s meant to do — surface up helpful content and bury content that’s not so helpful.`
Here’s a closer look at the three factors that might cause Google to revisit earlier content quality and ranking decisions:
- A page has weak or no originality
- It lacks firsthand expertise
- It offers low information-gain
Google’s not mad that you used AI. Google’s mad that you published bad content. And when you use automation to create lots of pages, none of which offer meaningful value to users, you’ve violated Google’s policy on scaled content abuse.
The damage lands hard. In her analysis of 220+ sites, Lily Ray found 39% lost 50% or more of their traffic, and 22% cratered past 75%. That’s not hypothetical. That’s real organic engines sputtering after months of artificial gains.
Here’s the deal: Google doesn’t really penalize AI-generated content. It’s all about the quality. And that quality starts with understanding what Google actually rewards, not what feels easiest to produce at scale.
What Truly Gets Rewarded by Google: Non-Commodity Content
Google might say it still wants to revisit AI content, but what people seem to miss is that Google isn’t penalizing you for using AI. It’s penalizing you for being lazy with it. Problem solved? Hardly.
Google indicated that the real goal is to get non-commodity content, and here’s what that means looking at…
- Unique. You bring something others don’t. A viewpoint competitors can’t replicate. Data only you have. An angle nobody else is covering. Commodity content recycles what everyone knows. Non-commodity content says “here’s what I found.”
- Specific. You’re referring to a specific context or scenario rather than an abstract take on a thing. For example, a content around “SEO best practices” is generic and not citable. But a case study that talks about how you resolved an SEO ranking crash in 48 hours is specific, citable and real.
- Authentic. It’s been done by you. First-hand experience. Original analysis. Proprietary data. The stuff that’s genuinely hard for competitors or AI to fake.
Officially, Google frames unique content as one that “brings a viewpoint or information others lack or can’t easily replicate,” specific content as one that “talks about a particular instance, not general rules,” and authentic content as one “demonstrates first-hand knowledge or expertise”.
Here’s why this matters for AI search: Google’s AI Overviews are now built to surface authentic voices and original sources specifically. The algorithm rewards content that’s hard to replicate—expertise, experience, authoritativeness, trustworthiness grounded in reality. That’s E-E-A-T, and it’s what actually sticks around when reassessment happens.
This is no longer a differentiator; it’s a standard. The single most important lever you’ve got for competing in AI search visibility is your non-commodity content. Everything else is noise.
The Growth Ceiling: Why AI Content Volume Caps SaaS Organic Growth
Here’s the SaaS playbook: scale AI content, watch organic traffic explode, and profit. The problem is, the growth cancels itself out.
AI Overviews decrease clicks to websites by 34.5% for informational queries
This is the price of putting out the same content as others in your industry. When you write commodity content, you’re up against thousands of pages all saying what’s yours is. So you’re not just competing against writers, you’re competing against clones.
And it gets worse: Strong SEO traffic is directly correlated with strong AI citation rates. Sites that already have strong rankings are cited more frequently in AI Overviews. Sites that don’t? They disappear. The data proves it. Over our analysis of 220+ sites, 54% saw 30%+ of their max organic traffic cut within a year. That’s more than a vibe—that’s a body count.
Differentiation helps you stand out as an expert, rather than another person in the field. Two components:
- Unique insight (proprietary research, original data, lived experience)
- Unique format (case studies, original analysis, frameworks competitors can’t replicate)
Multiply those together, and you’ve got content that AI Overviews actually cite. Everything else is just noise competing with itself.
But, how do you really build a content strategy that doesn’t crater six months in?
A Defensible Content Strategy: AI-Assisted, Not AI-Dependent
Everyone says “use AI strategically.” Almost nobody does. Here’s the gap.
| What AI Should Do | What AI Shouldn’t Do |
|---|---|
| Accelerate research & drafts | Replace your judgment |
| Surface trends & optimize prose | Become your source of truth |
| Speed up iteration | Generate final originality |
1. Use AI for scaffolding, not the walls.
AI excels at research sprints and draft acceleration. Let it surface trends. Let it tighten prose. What it can’t do is replace your judgment. If you’re publishing light-polished ChatGPT output, Google’s already counting the days until reassessment.
2. Anchor every piece to proprietary insight.
Your data is your moat — AI can’t manufacture it for competitors. Build around what you own: customer patterns, usage analytics, practitioner experience. That’s the difference between irreplaceable and forgettable.
3. Apply the UAS test: Unique. Specific. Authentic.
Does it say something others can’t replicate? Ground itself in real examples? Demonstrate first-hand knowledge? If it fails any of those, don’t publish.
Teams using AI publishing workflows—like ACME.BOT—dominate when differentiation is your editorial filter, not an afterthought.
4. Audit ruthlessly. Delete thin pages.
Scan existing AI-generated content for information gain. Thin pages signal vulnerability. Cut them or upgrade with original insight.
5. Track AI visibility in real time.
Some brands are tracking 36% visibility drops in weeks. If you’re seeing that pattern, adjust now—before the crash hits harder.
Differentiation isn’t optional. It’s your only defense.
The Bottom Line on AI Content Risk
Still think more content means more growth? Mass-produced AI content spikes, then crashes when Google realizes you’re publishing volume instead of value. AI itself isn’t the risk; it’s the risk of placing automation above judgement. You cannot risk not creating irreplaceable content that is informed by practitioner’s experience, customer data, proprietary insights that competitors can’t just prompt the AI to find. You can use AI to speed up your content creation, but not to replace your own thinking. If you do the first, your organic growth compounds. If you don’t, the crash is already in motion.