Following the exposure of some of the so-called "geo optimization of services" at the 3/15 party, the early concept of the industry was quickly labeled "to poison ai" "manipulate the large model". In a moment, geo went from "ai marketing new opportunities" to "new risks in the industry" and was even equivalent to grey operations。
We need to ask the following questions: what is geo? Is it a misused technology, or is it a misreading trend
A cognitive deviation amplified by short-term benefits
In the exposed cases, some agencies attempted to interfere with the results of the generation of ai through bulk production of content and the creation of false sources of information. This approach is not innovative, more like the transformation of the seo era's “black hat technology” - using the platform's rules loophole to make noise at the end of the information supply and mislead users。
According to the department of advertising and media economics at the chinese people's university school of journalism, this is a classic “new concept hijacked by old methods”: “geo is a completely new way of operating in the ai era, and the irregularities revealed are not really geo, but simply follow the old logic of seo and are exported through information piles
Professor and director, department of advertising and media economics, chinese people's university press school
“the problem is not ai nor geo, but rather the packaging of false business information as `objective answers'”。
Shunjie, general manager, omni performance, china
In other words, this is a typical “supply side fraud” rather than a deviation from the technology itself. He further added that “concealing false business information as `objective answers', at the current stage of partial ai-platform indexing and identification mechanisms, misleading consumer decisions through bulk, low-quality or even false content to infringe consumer rights. Some service providers, by capturing the brand/businesser's mentality of `complying with ai trends and pursuing short-term effects', have taken undue commercial advantage of the ai recommendation, which should have been objectively neutral, to be transformed into a `pay advertising slot'.”
More seriously, the grey operation was over-exposed and the whole geo industry image was defined in reverse. Thus, a question that should have discussed “how information is understood by machines” has been reduced to “how to manipulate machine responses”. This cognitive path itself is clearly misleading。
The essence of geo: not manipulation, but information matching
The logic of geo is not complicated from the very point of view of detaching market noise. The definition given by shunjie is that geo is essentially an “information fit” — it expresses real, consistent information in a way that is more in line with the ai semantic logic, making it easier for it to be understood, identified and quoted. The essence is to build a “trust chain” between the brand and the ai system。
This logic is similar to, but fundamentally different from, the early development paths of seos. In the age of the traditional search engines, the basic unit of access to information is the “link”, where users judge by clicking into content. In a generating ai environment, information is directly integrated into “the answer”, and users face a semantic compression and reorganization。
This means that the brand needs to deal with a target that has moved from “user browsing” to “machine understanding systems”。
Wang fei has put forward a deeper judgement on this: the core of brand competition is moving from being seen to being understood. If seo solves the question of whether it can be found, the question for geo is whether you are correctly understood when information is integrated into the answer. And that's why the simple equivalent of geo to "manipulating the ai answer" is a typical downside understanding。
“be referred to” to “understands”: the real shift in the problem
In the current market discussions, the most immediate anxiety of enterprises tends to focus on one question: how can ai mention me
But from wang fei's point of view, this is precisely a “lower issue”. She noted that the real core of geo was not “to be mentioned”, but “to be understood”. “remarked” is only a question of visibility that can be achieved in the short term by technological means and content optimization; while “understanded” is a structural issue, depending on whether the brand has a stable meaning system, a clear value anchor and a verifiable factual basis。
This distinction determines two completely different paths: one by increasing the probability of being quoted through content accumulation and semantic techniques, and the other by creating a long-term brand that allows machines to understand you in multiple circles in a sustained and stable way。
The former is more like “speculation”, while the latter is “capacity”。
In the operating logic of the ai system, this difference is further magnified. Generating models do not simply match keywords, but rely on a combination of semantic relationships, knowledge structure and multi-source information. A brand that lacks a stable structure, even if it is “referenced” at a particular node, is difficult to maintain consistency in the ongoing dialogue。
Thus, ai does not “codify” the brand, but only magnifies the integrity or absence of the existing structure。
One noteworthy phenomenon is that in the chinese market, geo tends to be quickly reduced to a "script, lay out, brush a sense of presence"。
According to shunjie, there are reasons for the technological phase behind this, as well as problems with industry mentality. “the industry needs to move as quickly as possible. Some service providers seek short-term results and rapid realization using short-term gaps in current ai index algorithms, neglecting the long-term mutual benefit relationship between brands and consumers. Both service providers and branders need to return to long-termism and insist on providing real, high-quality goods and information to consumers.”
Wang fei has a clearer judgment on this: if the brand does not have a stable meaning structure in the real world, its presence in the ai world is bound to be fragile. In other words, ai is not a shortcut to “brand construction”, but rather a more stringent “structure test”。
Ai search age: the brand is entering the machine portal
In a more macro-level perspective, geo is of concern because it is the back of a deeper change: the migration of the information portal. Over the past two decades, the user's information path has undergone multiple shifts — from search engines to electrical platforms to social media and short video. Today, generating ai is becoming a new core portal。
Shunjie mentioned that ai agent's monthly users have reached hundreds of millions of levels, and that more and more users are beginning to access information and assist decision-making through ai. This means that user attention and behaviour are being redistributed by ai. He further added that “especially for the younger generation, their growth is almost at the same time as ai, and the use of ai, trust ai, will become a long-term behavioural habit”
In this context, a key change facing the brand is that it is no longer just user-oriented but also machine-oriented。
That is why geo is becoming an emerging global capability. In overseas markets, commercialization exploration around ai search, generation recommendations has begun, and new assessment and optimization logic is being discussed by industry。
It should be emphasized, however, that geo is not an independent marketing tool but a structural upgrade of the entire digital marketing system。
Back to the practical level, how can enterprises cope with this change? Shunjie believes that the “content infrastructure” of the brand is the first thing to address is how real and effective information can be accurately understood and quoted by ai。
Specifically, it contains four key directions:
First, to understand the real problems raised by users in the ai scene, and to pre-position them in the product and service systems, from the real consumer demand perspective to avoid the content being removed from the actual use scenario。
Second, the structured expression of real information, the organization and organization of content based on the semantic logic of ai, helping ai to extract and integrate key information more efficiently, and enhancing the accuracy and visibility of brand information in generating responses。
Thirdly, the continuous output of high-priority heavy-trust sources, through the publication of high-quality content through authoritative media or credible endorsements, creates a stable and verifiable basis of trust。
Fourth, continuous monitoring and risk control, dynamic observation of the presentation of brand information in the ai semantic environment, ensuring content security and long-term stability。
The system essentially moves brand information from “dispatchable” to “understandable”。
For enterprises, geo reflects changes in consumer access to information, and brands need to adapt to such changes on a continuous basis rather than as a one-time technical tool or short-term tool. The reason is simple — the geo is not a technical upgrade, but a fundamental change in the way user information is obtained。
The vice president of google global advertising, dan taylor, also stressed that the core philosophy of google seo remains valid: focus on useful, reliable, people-centred content. Ai changed the way information is found, but did not change the definition of “quality content”, which remains the core criterion. Brands need to upgrade content strategies and increase input into high-quality pictures, videos and detailed product content。
Wang fei also believes that the truly high-level geo is not “feeding to the model”, but rather allowing the structure of meaning in the real world of the brand to be accurately read, understood and continuously used by the machine, which requires a brand presence structure. “the real geo, rather than creating an additional brand illusion for the ai world, completes the structural coherence between the real world and the ai world. That is, what businesses are going to do today is not to set aside the real brand and make an "ai-suitable brand version" on their own, but to ensure that brands exist internally in both worlds。
This strategy also presupposes that it must be based on truth and compliance. Otherwise, short-term “mentioned” may well be replaced by long-term “denied”。
Following the 315 events, the industry is revisiting the boundaries in this area. In response to a range of potential risks arising from the current disarray of services in the geo area and the absence of industry norms, the china advertising association launched a comprehensive effort to standardize the geo area in early march, in collaboration with a wide range of professional actors, including industry experts, academic institutions, brand enterprises and legal services. The formulation and preparation of the relevant standards are well under way. Zhang guohua, vice-president of the international advertising association (iaa) worldwide and president of the china advertising association, said: “in the generation of ai, advertising should not be short-sighted by irregular speculation, but must return to the cornerstone of `real trust'. The relevant standards in the area of geo are designed to address both the root causes and the root causes, to clarify the line of compliance and the code of conduct for the operation of the industry and to promote a healthy development pattern of `good money to drive away bad money'.”
In shunjie's view, the future will require two levels of consensus: a consensus at the value level — always centred on consumers, providing real, credible and verifiable information and adhering to long-termism; and a consensus at the rule level — including the traceability of sources of information, the normative aspects of content publication, and the capacity for scrutiny and technical governance on the side of the platform。
He added that “the three parties working together to complement each other is key to ensuring a healthy environment for content. A clear content production and auditing mechanism should be established between the agent and the brander, which needs to verify the true quality of the information before the content is produced, and the brand is responsible for verifying the accuracy and validity of the proxy content. Through joint clearances, the information ultimately provided to consumers is assured to be true and accurate. At the level of the platform, on the one hand, strengthening content auditing and pre-screening, and on the other hand, continuously optimizing indexing and identification mechanisms to enhance the accuracy and coverage of content governance at the technical level.”
Ai doesn't fake brands. It just magnifies reality
Back to the original question: was geo misread? The answer may be: it's not just the geo, it's the brand logic of the ai era。
When markets simplify it as a “manipulation tool”, what is overlooked is the deeper fact that ai does not create brands, but only understands brands. If the brand itself is not structured, ai will only expose problems more quickly; if the brand has a clear, real and stable meaning system, ai will become its amplifier。
As wang fei said, the real geo is not a brand fantasy for ai, but a brand-building match between the real world and the ai world。
This is probably what the industry needs to understand most after 315。
