The GEO industry is poised for a resurgence in value and a transformative renaissance.

©Original content from TMT Planet · Author｜Huang Yanhua

"Turning to Doubao for answers when encountering problems" has become a prevalent practice among many consumers. Amidst AI-generated responses, achieving brand visibility and product recommendations has emerged as a novel challenge for corporate marketing.

GEO (Generative Engine Optimization), a burgeoning technology, has thus come into the spotlight for numerous enterprises, being hailed as a potent tool for marketing breakthroughs in the AI era.

"TMT Planet" has observed that the rapid adoption of GEO, while introducing fresh perspectives to corporate brand marketing, has also spawned a plethora of chaotic situations.

According to a recent report by cnr.cn, amidst the swift expansion of the GEO market, issues such as "chaotic promises, erratic pricing, and disordered orders" have come to light.

The first batch of enterprises to "take the plunge" (a more idiomatic expression than "eat the crab," meaning to venture into something new and uncertain) not only failed to achieve instant success but instead found themselves as the inaugural batch of "victims."

01 Unfulfilled "Divine Promises"

"Rank in the top three on AI," "effective in seven days, refund if not" – such advertisements from GEO service providers are likely familiar to many. For numerous small and medium-sized enterprise owners feeling the heat in the AI era, these promises are particularly enticing.

For them, the advent of the AI era is inevitable. The pressing question is how their enterprises can capitalize on this trend to sustain their existence and even attract more users. Thus, investing a few thousand yuan monthly to feature their enterprise information in the recommendations of major AI platforms seems like a prudent move.

However, these enterprises were swiftly disillusioned by the rampant growth of the GEO market.

As reported by cnr.cn, an international logistics company established in 2024, noting the increasing number of clients mentioning being "recommended by AI," eagerly engaged a GEO service provider. The provider promised to "handle dozens of keywords and guarantee a top-three ranking across all platforms" at an "affordable" price.

Yet, a month later, the so-called ranking was merely a fleeting moment, quickly vanishing without a trace. Not only that, but to inflate the numbers, the service provider even churned out content in bulk without verification, fabricating false information such as the enterprise "having a factory spanning over 200 square meters" and "celebrating its tenth anniversary."

Subsequently, after cross-verification by the AI platform revealed significant discrepancies with the facts, the enterprise's hard-won exposure opportunities plummeted to zero.

Another moving company had an even more "bizarre" encounter. After paying the service fee, they registered multiple self-media accounts and continuously published articles as instructed. After nearly a month of toil, the so-called ranking and exposure remained nonexistent. Whenever they anxiously inquired about the results, the response was always "the technology is being refined" and "it will be ready soon." Ultimately, several thousand yuan in service fees were wasted, and the refund process was fraught with obstacles.

Product experts from 360 Zhijian GEO revealed the truth in an interview: Large models are not traditional search engines, and their answers do not adhere to a "fixed ranking." Search engines rely on fixed web crawling and sorting logic, whereas large models generate answers dynamically, influenced by multiple factors such as training corpora, real-time online search results, user question phrasing, and even scenario demands.

The promises made by those GEO service providers, such as "guaranteeing XX across the entire network," imply a stable appearance under all user queries, all phrasing methods, and all scenarios, which is technically almost an insurmountable task, with optimization costs far outweighing the benefits.

To swiftly achieve results and fulfill so-called "promises" to enterprises, some low-priced GEO service providers resort to non-compliant means such as "AI poisoning." By fabricating user reviews, assuming false expert identities, and forging data endorsements, they flood the common information sources of large models with false content, attempting to "teach" AI to lie.

Such behavior not only renders the enterprises' investments futile but may also lead to severe consequences such as brand demotion and restricted exposure due to the identification of non-compliant content by platforms. The repair process often spans several months.

02 Not Understanding AI's Operational Logic Equals "White Effort"

One significant reason why many enterprises fall prey to GEO service providers and have their expectations dashed is their lack of clarity regarding the underlying operational logic of large models, or they only possess a superficial understanding and attempt to apply SEO practices to GEO, believing that as long as there are sufficient keywords or information, they will be recommended by AI. However, there are fundamental disparities between the two.

The core of SEO revolves around keywords. The more and accurate the keywords, the higher the ranking in search engine results.

Whether GEO references an enterprise's materials does not solely hinge on "ranking." It places greater emphasis on the relevance of the content to the user's question, as well as the credibility of the source and whether it is supported by data or specific cases.

Therefore, content that is structurally clear, opinionated, supported by evidence, and from credible sources is more likely to be referenced by AI large models.

Another point is that each AI large model is an independent platform, and the AI large models created by different platforms will vary. For instance, just because Doubao references your enterprise's content does not mean Kimi will as well. Therefore, there is no one-size-fits-all GEO strategy that can guarantee stable results across all platforms.

If such a strategy exists, it is most likely a scam.

After comprehending the differences between the underlying operational logic of AI large models and SEO, the correct implementation of GEO becomes evident.

Firstly, your enterprise's brand needs to have a ubiquitous online presence, making it "searchable everywhere." Because AI large models are not real-time search engines; their information stems from vast training data and some real-time searches. If an enterprise's brand information is scarcely mentioned online, such as lacking a high-weight, high-inclusion-ranking official website, as well as third-party endorsements like encyclopedias, high-like content, and industry media, the probability of being referenced is significantly diminished.

Secondly, the problem direction you target should be sufficiently specific. If your enterprise provides legal services and you aspire for AI to mention you when answering "What are the best law firms in China?", your competitors are top-tier red-circle law firms like King & Wood Mallesons, JunHe, and Fangda Partners, which have been deeply "learned" by large models. The scope for GEO to make a difference here is exceedingly limited.

However, if you target a different question, such as "What are the top law firms specializing in economic cases?" or "What are the top law firms specializing in divorce litigation cases?", the competitive landscape undergoes a complete transformation.

Thirdly, output content from the perspective of AI large models. AI reads content differently from humans. Humans can discern the focus of a piece of content independently, but AI large models require you to explicitly present the focus or answer, preferably in bullet points or tables, so that AI can effortlessly grasp the key information.

Therefore, when creating brand communication content for GEO, lengthy brand stories or abstract concepts may not enable AI to extract effective information or determine if it is the answer to a user's question, making it difficult to be referenced and included.

From these three dimensions, the correct implementation of GEO is a systematic endeavor. It necessitates enterprises to first create high-quality content, structurally organize their brand information, and establish a knowledge system that AI can accurately identify and reference, laying a solid foundation for subsequent AI marketing efforts. It is definitely not something that "unscrupulous service providers" can accomplish by churning out a few fake articles in bulk.

03 It's Not Too Late to Rebuild Trust

It is noteworthy that the wild growth and frequent chaos in the GEO market have aroused strong vigilance among regulators and within the industry, which is a positive development for legitimate GEO service providers and the entire market.

In April this year, nearly 40 professional media outlets, industry organizations, universities, and technology enterprises jointly launched the "Responsible GEO Governance Initiative" in Beijing.

The initiative clearly states that the core of GEO is not to manipulate AI's answers but to convert content into knowledge that AI can comprehend through compliant means based on truth and value.

The initiative sets four red lines, including "opposing the fabrication of false information to 'poison' the internet, opposing the use of AI to generate low-quality content to 'pollute' the internet, opposing malicious attacks and unfair competition, and opposing the infringement of intellectual property rights."

Meanwhile, the evaluation work for the "Trustworthy Basic Requirements for Generative Engine Optimization (GEO)," promoted by institutions such as the Artificial Intelligence Research Institute of the China Academy of Information and Communications Technology, has also officially commenced, attempting to establish "rules" for GEO services from multiple dimensions such as management mechanisms, client material review, and the credibility of optimization methods.

The transition from wild growth to standardized development has become the only path for the healthy development of the GEO market and will also drive further expansion of the market size.

Guosen Securities predicts that by 2030, the global GEO market size will grow from $24 billion in 2026 to $100 billion.

For brand owners, ranking is illusory. Whether the content is true and worthy of being recommended by AI is what truly matters and is the correct way to utilize GEO.

When the sword of regulatory Damocles falls and industry self-regulation consensus forms, the GEO industry will undergo a resurgence in value and a transformative renaissance.

*The featured image in the article is from the official website of Weimob Xingqi.