In the domestic Geo industry, there is finally an "operation manual".

This year's March 15th party exposed the black industry chain of Geo's fake corpus-polluted AI platform, and the chaos of fictional products misleading users completely surfaced.

Only half a month after the exposure, the China Commercial Advertising Association formulated the project initiation standard. After nearly half a year of drafting, soliciting comments and revising, on September 14, 2026, the five-part Geo group standard was centrally announced on the national group standard information platform.

Hangzhou Gelix Artificial Intelligence Co., Ltd.As a core participant unit, participate in the development of this standard throughout the process, and deeply participate in the compilation of the core provisions of Part 4 (Trusted Corpus Specifications) and Part 5 (Service Process Compliance and Safety Requirements).

The following is the core content, industry significance and business landing method for you to disassemble the entire set of standards in an all-round way.

01 | Geo Standards Overview

At present, two sets of Geo related group standards have been released in China:

Organizer	Standard Name	Focus	Implementation date
China Federation of News Technology Workers	Generative Engine Optimization (GEO) Trusted Information Dissemination and Information Ecology Governance Specification	Credible Information Dissemination and Ecological Governance	August 25, 2026
China Business Advertising Association	Generative Engine Optimization (GEO) Community Standards (as interpreted in this article)	Geo service full process specification, 5 parts	September 14, 2026

This article focuses on the five-part standard published by the China Business Advertising Association, from the definition of terms to the full coverage of service delivery:

PART	Standard Name	Core content
Section 1:	General Comments on Terms and Definitions	Harmonize Geo definitions, boundaries with SEO/AEO, performance metrics, compliance bottom line
Part 2	Service Provider Evaluation Specifications	Set up access and grading from a three-dimensional assessment of technology, services, and compliance
SECTION 3.	Specifications for Pricing, Evaluation and Measurement	Clarify pricing units and modes, establish uniform performance indicators and measurement methods
PART 4	Trusted Corpus Specification	Standardize corpus grading, source standards, quality requirements, and traceability mechanisms
SECTION 5.	Service Process Compliance and Safety Requirements	Standardize the operation of the whole process, draw three safety red lines, and establish a violation handling mechanism

In a nutshell:Two sets of standards in parallel, this article focuses on the five parts of the China Business Advertising Association's full process standards, from terminology to delivery one-stop coverage.

02 | General Discussion of Terms and Definitions: Unifying Terminology, Ending "One Person, One Word"

Before the introduction of the standard, the definition of Geo was confusing - some said it was an extension of SEO, some said it was an AI corpus operation, and some organizations classified all kinds of services into Geo. Conceptual confusion makes it impossible for customers to judge what they are buying at all.

Part 1 unifies four core terminology categories:

Geo definition and boundaries with SEO, AEO

Clarify the definition of Geo (Generative Engine Optimization) and distinguish it from SEO (Search Engine Optimization) and AEO (Answer Engine Optimization).
Corpus, knowledge base, optimizing entities, and other foundational concepts

Define corpus (information content referenced by AI learning), knowledge base (structured knowledge collection), optimized entities (optimized brands, products, people, etc.).
Effective indicators such as visibility, adoption rate, ranking changes, etc.

Establish unified evaluation indicators such as AI platform visibility (the number of platforms where information appears), content adoption rate (the proportion of target content adopted in AI responses), and recommended ranking changes.
Compliance & Ethics Bottom Line

Clearly prohibit fake content, data pollution, manipulation of AI recommendations, etc., and draw a bottom line for data security and personal privacy protection.

In a nutshell:Uniform terminology is the first step in industry normalization, and customers are no longer dazed by the concept of "one person, one story".

03 | Service Provider Evaluation Code: Three Dimensions at Your Fingertips

After the exposure of March 15, the problem of industry access was highlighted - some institutions claimed to provide GEO services by purchasing only one set of AI tools, and there was no self-developed technology or standard process. Customers face uneven service providers and lack a unified basis for judgment.

This part establishes a three-dimensional evaluation system for technical capabilities, service capabilities, and compliance capabilities:

Assessment dimensions	Inspection content	Parameters
Technical Proficiency	Self-developed algorithms, multi-platform adaptation, corpus processing	Whether to have self-developed core algorithms; the number of AI platforms covered; the depth of corpus processing technology
Capacity of service	Standardized processes, project management, impact monitoring	Whether there is a standardized service process; whether the project management mechanism is perfect; effect monitoring and feedback system
Compliance Capabilities	Data security, content review, risk disposal	Data security management system; content audit mechanism; risk identification and emergency response capabilities

Access and grading mechanisms:

mechanisms	Details
Entry Barriers	Reach the basic requirements of 3D evaluation before entering the GEO service market
Rating	The service provider is divided into different levels according to the overall score, and the customer selects them according to their needs.
DYNAMIC ADJUSTMENTS	Review regularly to adjust tiers based on quality of service, customer feedback, and compliance

This system allows customers to choose service providers with evidence, and no longer only listen to verbal promises from sales.

In a nutshell:Technology, service, compliance three-dimensional evaluation + access grading, choose service providers no longer rely on sales to beat the chest.

04 | Code of Pricing, Evaluation and Measurement: Pricing and Effects, No Longer "in Conscience"

Previously, the GEO industry was confused in terms of pricing - by keywords, by month/quarter, by AI referrals, directly packaged prices… the same service could be quoted several times differently by different agencies. Effect evaluation also lacks a unified caliber, and the word "good effect" can cope with the customer.

This section addresses two core issues:

Three pricing models

Pricing Mode	Billing Basis	Applicable Scenario
By Corpus Level	Amount of content to process and optimize	Projects with large content volumes that need to be optimized in bulk
By Optimization Period	Service Duration (MM/Q/YY)	Brands that need long-term continuous optimization
By Performance Indicator	Metrics such as visibility, adoption, etc.	Results-oriented projects with clear KPIs

Three Uniform Performance Indicators

Performance Indicators	Definition	Measurement method
AI Platform Visibility	How the brand message appears in AI platform responses	Number of platforms covered by statistics and frequency of occurrence
Content Adoption Rate	Proportion of AI responses that adopt the target content	Proportion of AI responses sampled that include brand target information
Referral Rank Change	Changes in the ranking of brands in the AI recommendation list	Compare the position of brands in AI recommendations before and after optimization

Key requirements: The service provider must disclose the data caliber, measurement method and statistical cycle to the customer, and the customer has the right to verify the data source. Thoroughly say goodbye to vague statements such as "works better".

In a nutshell:Pricing has three modes, and the effect has three indicators. Where the money is spent and how the effect is, all of which can be verified.

05 | "Trusted Corpus Specification": Controlling the Content of AI "Eating" from the Source

The accuracy of the AI response depends on the credibility of the input corpus. False assessment, fabricated data input AI, the output may deviate from the truth. 3 · 15 "Apollo-9" event is the recommendation distortion caused by false corpus.

This part specifies corpus from four dimensions: grading, source, quality, and traceability:

Dimension	Core Requirements	Specific instructions
Levels	Rating corpus by source and credibility	High-level corpus is preferred by AI, and low-level corpus requires additional verification
Source	Clear criteria for official publication, authoritative media, verified third parties	Government/agency officials are at the highest level, followed by authoritative media; anonymous or unverifiable content must not be entered as corpus
VBR Quality	Content must be true, accurate, and complete, and must not be fictitious or misleading	Corpus must be fact-checked, data must have a clear source; fabrication of products, fictional assessments are prohibited
Retrospection	Each corpus comes with metadata such as source, time, author, etc.	Establish a metadata recording mechanism to quickly locate and deactivate responsible subjects when problems arise

Three levels ensure that the corpus is trustworthy:

🔍 Pre-warehousing verification: new corpus must go through a credibility assessment before entering the knowledge base
🔄 Regular review: Regular review of incoming corpus, timely update or deactivation of invalid sources or outdated content
⚠️ Anomaly monitoring: monitoring AI responses for false or misleading content, finding problems back to the corresponding corpus and disposing of them

Analogy: Just like food and beverage ingredients procurement standards - the ingredients entering the kitchen are prioritized for freshness and quality, and the quality is controlled from the source.

In a nutshell:Corpus grading + source verification + three levels, control AI input from the source, false content can not enter.

06 | Service Process Compliance and Safety Requirements: Five-Step Specification with Three Red Lines

Geo services involve customer business information, data materials, content production and multi-platform operations. Mismanagement of any link can lead to data breaches or content violations.

Five-step service process:

Step	Contents	Key Requirements
① Customer access	Assess customer qualifications and needs	Non-compliance request not accepted
② Material review	Review customer profile for authenticity and compliance	Establish data hierarchy protection
③ Content production	Establish a content review mechanism	Complies with corpus specifications, leaving traces throughout the operation
④ Optimized implementation	Multi-platform optimization according to standard process	Violation tools and manipulations such as fake content and data contamination are prohibited
⑤ Delivery acceptance	Provide traceability report	Transparent data caliber, customer acceptance against standards

Three safety red lines 🚫

New date	Acts not Allowed
Information and Content Security	Prohibit production and dissemination of false content
Data and Personal Information Security	Prevent disclosure of customer data and user privacy
Technology & Tool Safety	Prohibit the use of offending tools and manipulation of AI-recommended means

Violations are dealt with by grade, and serious violations face industry discipline. Analogy: Just like dining kitchen specifications - cooked separately, tableware disinfection, personnel hygiene, there are standards for each step.

In a nutshell:The five-step process + three red lines, from the pick-up to the delivery of each link, there are standards to follow, and violations are punished.

07｜GerexWhy you can participate in core drafting

Of the 41 participating units,GerexParticipated in the content drafting of Part 4 (Trusted Corpus Specification) and Part 5 (Service Process Compliance and Safety Requirements).

Cf. Part 4: Technical Accumulation of Corpus Governance

GerexEarlier, I focused on the Geo track and built a self-developed technology system around AI question-answering semantic logic:

Seven self-developed core algorithms
Semantic distillation, knowledge base construction, entity recognition, credibility assessment, multi-platform adaptation, effect monitoring, risk warning

Service adapts to nine major AI platforms
Beanbag, DeepSeek, Kimi, Tongyi Qianqian, Wenxin Yi, Tencent Yuanbao, ChatGPT, Gemini, Claude

As of August 2026, it has served more than 300 enterprises. The first-line corpus processing experience provides a server-side reference for the Part 4 standard.

Reference Part 5: Practical precipitation of service processes

GerexStandardized service process: Building a → knowledge base for needs diagnosis → Content optimization → Multi-platform → distribution monitoring

Supporting three mechanisms of data security, content compliance, and risk warning, clarifying operational traces, data desensitization, and red line requirements - these practices provide a reference framework for Part 5 standards.

Industry Qualifications and Standards Participation

Director of China Advertising Association
Governing Unit of China Business Advertising Association (AI Marketing Application Working Committee Member Unit)
In addition to the Geo series of standards of the China Business Advertising Association, it is also the drafting unit of the China Federation of Journalism and Technology Workers T/capt 026-2026

One Sentence Summary：Seven algorithms + nine platforms + 300 enterprise experience, yesGerexThe bottom of the two parts of the core of the participating editions.

08 | Standard landing, how to use it for enterprises

Impact of Standards on Three Parties

Objects	Implications
Service Provider License	Compliance capabilities have a unified assessment standard, making it easier for institutions with technical and service capabilities to be recognized; the living space of institutions that rely on non-compliance and low-cost competition is compressed
BUSINESS CLIENTS	There is a clear basis for selecting the service provider and the inspection effect, which can be examined against the standard terms
AI User	After the corpus is trusted and the process is compliant, the probability of false recommendations in AI responses is expected to decrease

Enterprises choose Geo service providers and directly ask these three questions ✅

First question (corresponding to section 4):
Are the corpus graded and source-checked according to standards? Can you provide a source description and a trust level document?
Second question (corresponding to section 5):
Are customer access, data protection and risk handling mechanisms in place? Is it possible to provide a full traceability of the process?
Third question (corresponding to section 3):
Are standard uniform measurement methods used for effect evaluation? Is the data caliber publicly verifiable?

Gerex's landing motion
After the release of the standard,GerexThe standard content has been disassembled into a project delivery checklist: the corpus source can be traced, the service link is recorded, and the effect data caliber is transparent.

In a nutshell:Choose a service provider and ask three questions. Compare the standard terms and conditions, do not step on the pit and do not spend money.

About Gelix

New Galaxy AI

The first batch of Geo service providers in ChinaSeven self-developed core algorithmsCover nine AI platforms300 + enterprise customersRenewal Rate 97%

Two large group standard drafting units$10M Financing

Hangzhou Gelix Artificial Intelligence Co., Ltd.(New Galaxy AI) is the first AI technology service provider in China to lay out the Geo circuit.

Core capabilities:Seven self-developed core algorithms, five software copyrights, and self-developed prism monitoring engines cover nine mainstream AI platforms at home and abroad, and have tens of millions and millions of Geo project optimization experience.

Industry Qualifications:The governing unit of the China Advertising Association and the governing unit of the China Commercial Advertising Association (member of the AI Marketing Application Work Committee) participated in the preparation of the Generative Engine Optimization (GEO) Group Standard and the drafting unit of the Generative Engine Optimization (GEO) Trusted Information Communication and Information Ecology Governance Code (T/capt 026-2026) of the China Federation of News Technology Workers.

Serving Size：Cumulatively, it serves more than 300 enterprises, covering 28 industries such as medical, cultural and tourism, technology, finance, manufacturing, and urban brands.

Financing:Finished $10 million financing in January 2026.

The domestic Geo industry welcomes the official release of the operation manual | New Galaxy AI in-depth participation in the five Geo group standards