At the 2026 CCTV 3·15 Gala, a shocking tactic called “poisoning AI” went viral overnight.
The report uncovered a gray industry chain around GEO (Generative Engine Optimization), exposing a hidden side of AI marketing. With nothing but a completely fictional product, reporters worked with service providers to mass-produce fake reviews. In just 72 hours, several major AI models started recommending this non-existent product as a top choice.
In the age of AI, a trust crisis has been revealed in an absurd way—and is now accelerating in an equally absurd way.
Ironically, this exposure didn’t slow the GEO industry down. Instead, it gave this once niche concept a nationwide spotlight. After “3·15,” many service providers actually saw more inquiries and orders, not fewer.
On one side, regulators are calling out the chaos. On the other, market demand keeps rising.
So what exactly is GEO? A new marketing frontier in the AI era—or just the same old traffic scam in a new disguise?
01 | GEO: Quietly “Poisoning” AI Models?
GEO stands for Generative Engine Optimization.
In the past, companies relied on SEO (Search Engine Optimization) to rank higher on platforms like Google or Baidu. Now in the AI era, that logic has evolved into GEO—its goal is to make AI models prioritize and recommend specific brands when answering user questions.
“We don’t have complete conversion data yet, but you can clearly feel the market heating up,” said Liu Jie. Less than two weeks after the 3·15 Gala, inquiries for his company’s GEO services surged.
Liu and his team have been in the marketing space since 2010. In the PC era, they did SEO and SEM. In the mobile era, they built H5 campaigns and mini programs. By 2025, as AI models became mainstream, user behavior shifted from “searching websites” to “asking AI.” That’s when they spotted the GEO opportunity.
Many people now equate GEO with “AI poisoning” or fake marketing. But practitioners see it differently: GEO is just a neutral tool—an extension of SEO in the AI age, not some brand-new scam.
As Liu explains, the core of GEO is simple: translate a product’s strengths into a format that AI can understand and trust.
Why is that necessary?
Because humans can accept vague claims like “better performance” or “great experience.” AI can’t. It prefers structured, verifiable, and authoritative information.
For example:
- If you say a product is faster, you need to specify how much faster.
- If you claim lower energy consumption, you need exact percentages.
- Ideally, you back it up with patents, awards, or third-party reports.
Another industry insider, Xia Xia, who has worked in Google SEO for years, agrees that GEO isn’t technically groundbreaking.
“In our industry, GEO is usually just an add-on service. Almost nobody does GEO alone. If someone claims they only do GEO, that’s a red flag—they’re probably just trying to make quick money.”
Xia Xia came from the export business, where SEO is essential. In his view, GEO is built entirely on SEO foundations. Both rely on high-quality content. The difference is:
- SEO also cares about site speed, mobile optimization, backlinks, etc.
- GEO focuses almost entirely on content itself.
“As long as your content is high-quality and authoritative enough to be picked up by AI, you have a chance to be recommended.”
“For most people, you can think of GEO as just a branch of SEO. There’s no revolutionary technology behind it.”
02 | Not a Tech Barrier—The Real Challenge Is Content
Despite the negative exposure, GEO didn’t take a hit. In fact, awareness drove even more business.
According to Liu, the surge in demand comes from three main factors:
- Many companies are hearing about GEO for the first time and realizing AI is a new marketing channel.
- GEO is still in its early “gold rush” phase—low barriers, quick results.
- Defensive demand is rising. Companies worry competitors may spread false information, so they want to publish accurate content first and “claim” AI mindshare.
Meanwhile, Xia Xia’s business—focused on Google SEO for export-oriented factories—hasn’t been affected much.
“My clients are mostly B2B manufacturers—industrial equipment, chemicals, metal casting. Their priority is still getting leads through Google. Only a few ask how to get recommended by ChatGPT, and it’s not mainstream yet.”
Why hasn’t GEO exploded in overseas markets?
Because leading global AI models evolve faster—and are better at detecting fake content.
“Some Chinese service providers think the more detailed the content, the easier AI will trust it. That doesn’t work with models like ChatGPT or Gemini. They care much more about credibility and sources, not just details.”
Both experts agree: the biggest misunderstanding about GEO is thinking it’s a highly technical field.
It’s not.
At its core, GEO is a content strategy problem, not a technical one.
Liu’s company charges around $1,300–$1,500 per project. The focus isn’t on a few keywords, but on covering multiple user scenarios with optimized content. Performance is measured by:
- Mention rate
- Recommendation rate
- Top-ranking rate
Xia Xia, on the other hand, charges around $14,000 annually for SEO services. His process is much heavier:
- Build a proper website
- Deeply understand the client’s products and capabilities
- Map out customer pain points
- Create tailored content to address them
“It can’t be mass-produced. Otherwise, search engines will penalize you.”
Low-quality GEO or SEO services usually rely on AI to generate massive amounts of spam content and distribute it everywhere.
“This model has existed for over a decade,” Xia Xia said. “It may boost traffic short-term, but within 6–12 months, the site gets penalized and basically dies.”
03 | GEO Is Just Repeating SEO’s History
What’s happening in GEO today has already played out in SEO over the past 20 years.
When search engines first took off, SEO became essential for online marketing. The industry quickly turned into a battle between “black hat” and “white hat” tactics:
- Keyword stuffing
- Spam websites
- Fake clicks
- Buying backlinks
Many providers made quick money through shortcuts, leaving the industry full of chaos.
But over time, algorithms improved, and regulations tightened. Laws like the Advertising Law and Anti-Unfair Competition Law reduced the space for manipulation. Opportunistic players gradually disappeared.
Now, GEO is following the exact same path.
The fake content and “AI poisoning” exposed by 3·15 are just symptoms of an early-stage, fast-growing industry.
Xia Xia put it bluntly:
“This is just the same old trick—wearing an AI mask. Even today, many SEO providers in China still do this. It’s all about making quick money.”
But a clear signal is emerging:
Those who rely on “poisoning” tactics won’t last.
On one hand, regulation is inevitable. Standards will be established.
On the other, AI platforms have strong incentives to fight manipulation. If users realize AI recommendations are full of fake information, trust collapses—and so does the product.
For AI companies, ensuring reliable sources isn’t optional—it’s survival.
So how should businesses approach GEO in the AI era?
Both experts give the same answer: Forget shortcuts. Stick to real, compliant content.
Liu believes the real edge lies in execution:
- Deeply understand the product
- Identify true differentiation
- Support claims with credible evidence
- Adapt to AI platform rules
Xia Xia is even more direct:
“SEO is content marketing. GEO is the same. Whether you’re writing for humans or AI, the goal is the same—answer real questions and meet real needs.”
“All those ‘optimization tricks’? They’re just icing on the cake. Without real, high-quality content, everything else is meaningless.”
From search engines to social platforms to AI models, the channels keep changing—but the core logic hasn’t:
Trust is built on truth.
The 3·15 exposure isn’t the end of GEO. It’s the beginning of its normalization.
And the best optimization has never been about gaming the algorithm—
It’s about becoming worth recommending.