This content is excerpted from a dialogue between Monica and Uncle Zou on April 15, 2026.
Guest Introduction
Uncle Zou
CEO of Bangbang Growth & WhatGEO
Author of books including ToB Growth Operations, Global Expansion Without Limits, and AIO & GEO Practice: How to Become the Answer to AI Recommendations.
Monica
Let me give a brief introduction to Uncle Zou. He has been deeply engaged in the marketing industry for years, serving many large ToB enterprises with extensive practical experience. I contacted him recently and learned he is releasing a new book on GEO. Therefore, we are honored to invite him to share insights with us today.
Uncle Zou
Hello everyone. I’d call myself a seasoned veteran committed to the marketing industry long-term. I previously oversaw brand and marketing operations at SaaS companies including Chuanglan Cloud, Zhiyuanyuan Cloud, and Full-Stack BI. Frankly speaking, the SaaS industry has faced tremendous headwinds in the past two years, with part of its business being replaced by AI.
I started my entrepreneurial journey a few years ago, staying consistent with my work in the international business division of Chuanglan a decade ago. I focus on overseas marketing, providing full-suite solutions mainly for manufacturing and tech enterprises. A clear observation is that standalone overseas marketing has become extremely difficult nowadays. Hence, we only partner with enterprises with long-term vision that demand one-stop integrated marketing solutions.
The concept of GEO first emerged among enterprises in late 2023 to early 2024. To be honest, I had never heard the term “GEO” before August 2025. It was not called GEO overseas back then; the mainstream terms were AI Optimization or AI Visibility.
In early 2025, my partners and I independently served over a dozen enterprises and began writing this new book. We even revised the book title before the manuscript submission in August and September. The original title did not include “GEO”, but we added this core keyword later as it had become the industry standard term in China. All marketers understand how critical core keywords are. Currently, the new book is available for pre-sale online, and physical copies will be available next week.
The book is titled AIO & GEO Practice, with the subtitle How to Become the Answer to AI Recommendations. I think the subtitle is straightforward and serves as a clear action guide.
I’ve had an increasingly strong realization lately: although AI traffic overall is not yet enormous, the first user touchpoint for many brands—especially emerging brands in recent years—is very likely through AI recommendations. Paid advertising remains costly. As a user with precise demands, I will most likely turn to AI for answers first. If your brand appears right then, it becomes the most efficient first engagement between you and the user.
We also have colleagues from the Qianwen team in today’s live stream, and we are their external commercial partner. One of their cross-border e-commerce clients had never done SEO or GEO before, relying purely on platform e-commerce tactics. When they consulted us about launching GEO, my response was that such enterprises do not need standalone GEO deployment. Instead, they should integrate branding, content, SEO and social media—a proper path toward brand transformation.
Cognitive Foundation: Understand AI’s Logic Before Crafting Strategies
1. The Real Stage & Traffic Value of AI Marketing
People often ask how mature AI marketing has become. Frankly, China leads the world in the popularization of AI applications. Not only professionals but also the elderly and children use Doubao in daily life, making it indispensable. This is driven by the massive user base accumulated by consumer-facing AI products such as Doubao, Yuanbao and DeepSeek.
Chinese users show extremely high acceptance of AI, while overseas adoption is far less widespread than we assume. Various industry reports indicate AI-driven traffic currently accounts for only about 1% of overall global traffic.
However, there is a remarkable advantage: traffic generated by Large Language Models (LLMs) boasts a conversion rate nearly three times higher than direct access and traditional search traffic. The logic is simple. Any question or demand description users input into AI functions as a Prompt. Thus, users asking questions on AI platforms are high-intent, precise traffic.
2. The Essence of GEO: Not Magic, But Science
You may have heard numerous interpretations of AI recommendation logic. Fundamentally, if you consistently produce high-quality content and develop overseas markets sincerely—this applies both domestically and globally—your brand will eventually appear in AI answers once your content and SEO efforts reach a certain scale.
Why do we still need to study GEO exclusively? Because with a GEO mindset, you will better understand AI’s underlying mechanisms when conducting SEO and content marketing, enabling more precise and efficient operations. Mastering 100 conventional SEO tasks does not automatically grant you GEO capabilities. We must focus on the differences between SEO and GEO.
A key insight: we learn Google’s algorithms and rules first when doing SEO, not marketing theories. Similarly, to leverage AI tools such as GPT and Claude, we must first master AI’s inherent rules. My new book dedicates extensive content to explaining the principles of LLMs and RAG, which is rare among similar publications on the market.
Before practicing GEO, ask yourself one question: Is our core goal to have our brands and products mentioned by AI when users raise relevant questions? That is essentially what GEO is. To achieve this, we must first understand a prerequisite: How do large language models generate content? Without this foundational knowledge, you are easily misled.
I often hear potential clients claim: “We know a company that can reverse-engineer AI algorithms and figure out how GPT generates content.” My reply is straightforward: “It’s impossible. If they had such capability, they would build their own LLM instead of offering GEO services.” Others claim they have technology to embed content into LLM training processes.
My perspective: As marketers or non-technical entrepreneurs, we do not need to delve too deeply into AI algorithms, but we must learn and grasp underlying technical principles. Only then can we formulate accurate and reliable GEO strategies.
3. Decode AI’s Core Logic: Key Essentials You Must Know
RAG Technology: The Core Gateway to Influence AI
LLMs have inherent flaws including hallucinations, poor numerical reasoning and lack of real-world comprehension. To solve inaccurate content output and failure to access up-to-date information, developers adopted a powerful technology: RAG (Retrieval Augmented Generation).
Do not be intimidated by the technical term; simply understand it as web-connected search. When AI cannot rely solely on its pre-trained internal knowledge base, it accesses real-time internet data for reference.
This brings a vital truth for marketers: nearly all our GEO operations can only affect AI’s real-time web retrieval phase. We cannot alter fixed pre-trained model data.
Let me share a recent case. One of my pharmaceutical clients found that certain AI models directly gave answers without accessing the internet when users asked about medications for specific diseases. The client wondered if optimization was still possible. I replied it was not feasible. Some AI models skip web retrieval if their pre-trained database already contains complete, stable answers to certain questions.
Every web crawl by AI acts as a real-time evaluation of your content and brand. Overseas markets set extremely high standards for content authority and structural integrity, which we must prioritize.
Part Two: Hands-On Practice – A Minimal GEO Strategy Anyone Can Deploy
With foundational principles covered, let’s move to practical implementation. I will walk you through building a minimum viable GEO strategy step by step in 20 minutes. What I share today is not the full scope of GEO, but an MVP model perfectly suited for small and medium enterprises with limited resources and no budget for third-party service providers.
The strategy boils down to three core steps, defined as the RDIM Model:
Reverse, Deep Learn, Imitate & Surpass
Nowadays, corporate operation and growth roles account for a much higher staffing ratio than before. In earlier years, few companies had dedicated operation staff; even those assigned related tasks worked part-time, spending only 1/4 or 1/5 of their time on operations. Over 90% of employees were in product and R&D teams back then, while the ratio has now balanced to nearly 1:1.
Step 1: Reverse – Map Out Industry Competitors
Action Guidelines
Establish a tiered question framework: For most brands, you can ignore Tier-1 questions (e.g. "what is the best..."), as answers are always dominated by industry leaders with little relevance to mid-tier and small brands. We need to disassemble user questions strategically.
Interactive Exercise: Take your product or brand and design three tiers of questions:
Tier 1: the best + [your product/service/brand]
Tier 2: the best value + [your product]
Tier 3: Add country/region restrictions based on Tier 2, e.g. the best value [your product] in Germany
Why tiered questioning? Broad generic questions will always exclude small and medium brands. Such enterprises should focus on Tier 3 to Tier 5 long-tail questions to capture niche traffic.
Question & Record: Open three independent windows on ChatGPT or Gemini to avoid context interference, and input the three tiered questions. Focus on three key observations from the answers: Which brands are mentioned? Which are highly recommended? Does your brand appear?
If your brand is already referenced by AI without deliberate GEO efforts, this hints at a deeper concept I will elaborate on later: GEO as an independent service may not even exist. After this step with no extra tools required, you gain your first key insight: the landscape of AI-recommended brands in your industry.
Step 2: Deep Learn – Analyze Top-Performing Benchmarks
In most cases, your brand will not appear in AI answers after Step 1—and this is exactly why we need GEO.
Action Guidelines
Disassemble core sources and content structureOpen the cited source links from AI’s answers. Take our AI video product client as an example: Runway is frequently recommended, with PC Magazine as one of its core authoritative sources.
Sample source article title: Best AI value — our generation tests...
The article features structured parameter comparison tables, editorial scoring, and in-depth pros & cons analysis.
You only need to analyze 3 to 5 such articles to grasp the standard framework. You will gain your second key insight: what types of authoritative sources, title structures and content frameworks make industry leaders stand out in AI recommendations. You will learn both what to write and how to structure contentto align with AI preferences.
Map citation source channelsIdentify where high-quality referenced content is published and which platforms serve as mainstream authoritative sources in your industry. For cloud vendors, official blogs of Cloudflare, AWS and Google Cloud dominate top-tier sources—platforms inaccessible for most SMEs. Gartner rankings also carry high authority but are unaffordable for small businesses, often costing over one million to secure a listing.
Audience Question: How to analyze the top 100 authoritative sources?
Uncle Zou’s Response: Manual analysis works, but a smarter AI-era method is to feed all cited links from AI answers to another LLM, and let it sort and summarize the top source list. You will gain another critical insight afterward: official websites are irreplaceably important. In overseas markets, brands cannot operate effectively without an official site—it is your only free, fully controlled voice platform.
Step 3: Imitate & Surpass – Evolve from Follower to Leader
After disassembling questions, sources and content structures, SMEs face a realistic challenge: most authoritative platforms are inaccessible for content publishing.
Action Guidelines
Identify accessible publishing channelsScreen the ranked top 100 source list to find blogs and websites open to external contributors. Additionally, leverage high-authority open platforms:
Free publishing: Reddit, LinkedIn
Video content: YouTube
Guest posting & Q&A: Medium, Quora
Social platforms: Facebook, Yahoo
For content creation guidance: Analyze the tone and format of content cited by high-authority platforms (marketing-oriented or educational-oriented), then replicate their framework. When lacking clear direction, benchmark industry top performers first.
An interesting finding: Long-form Articles on LinkedIn are far more likely to be cited by AI than short Posts, thanks to higher information density that aligns with Google ecosystem’s definition of valuable content.
Achieve breakthrough via differentiationImitation is only the starting point. True industry breakthrough requires defining your product positioning and competitive differentiation.
For example, smartphone brands cannot compete with Apple, Samsung or Huawei for Tier-1 and Tier-2 rankings for "best smartphones". But if you specialize in elderly phones, AI may recommend your brand when users ask for "the best elderly phone", instead of mainstream giants. Similarly, if you launch rugged phones with 20,000mAh batteries and underwater photography capabilities, you can outperform mainstream brands for queries about "best rugged phones".
GEO forces brands to rethink product value propositions beyond mere algorithm optimization.
Build enterprise knowledge graphs & semantic relevanceSystematically sort out your core brand assets: brand name, product lines, technologies, materials, industry standards and competitive advantages. Integrate these keywords into your content consistently to build semantic connections.
In simple terms: AI initially has no recognition of "Monica", but establishes relevance through repeated exposure to "Monica’s Global Expansion Talks". Similarly, repeatedly mention your brand and product keywords across multiple articles to narrow their vector space distance in LLM models.
Final Step: Monitoring & Iteration – Measure Your Outcomes
Complete the above operations and establish a performance baseline. Record AI answer results for your question matrix as of April 15, 2026. Reassess changes after one month of optimization.
The moment you see your self-built content cited by AI, transforming your brand from invisible to visible in AI recommendations, you will experience the classic aha moment of GEO practice.
Important Reminder: Avoid tools like Profound. First, they are overpriced; second, their monitoring data is inaccurate. Most such tools rely on API interfaces, whose raw output differs drastically from the polished answers users see on ChatGPT/Gemini web and app versions. Using developer-side raw data to optimize user-end presentation is counterproductive.
Part Three: Ultimate Strategic Thinking – The Essence of GEO Lies in Brand & Content Marketing
Beyond tactical skills, I’d like to share deeper strategic reflections that I’ve been verifying and questioning throughout this year.
Does Independent GEO Truly Exist? An Interesting Paradox
Our analysis of extensive industry reports reveals a pattern: Broad generic queries such as "best AI video tools" always feature top players like Runway and HeyGen; queries for "best energy storage brands" rank Huawei and Sungrow at the top.
Did these brands deploy dedicated GEO strategies? Almost certainly not.
Why do they dominate AI rankings effortlessly, while many brands investing heavily in standalone GEO see little return?
My conclusion: Independent, pure-play GEO may not exist at all. What we practice as "GEO operations" is essentially making up for past shortcomings in branding, content marketing, SEO and PR.
The superficial logic of AI recommendation lies in content marketing, while the underlying core is consensus building through multi-party cross-verification. Brand credibility is not self-proclaimed; it is validated by third-party reviewers, industry peers and authoritative media outlets.
The New Paradigm of Branding & Marketing in the AI Era
Therefore, GEO should not be treated as an isolated task, but a new mindset integrated into all marketing workflows. It drives fundamental shifts in brand and marketing strategies:
Brand Strategy: Shift from brand image building to establishing AI-verifiable brand entities (e.g. Wikipedia entries, Trustpilot profiles, standardized Schema markup).
Content Strategy: Shift from traffic acquisition to creating citeable source content and modular answer frameworks.
Communication Strategy: Shift from media exposure chasing to industry consensus shaping.
Data Analysis: Build new KPI systems including brand visibility rate, top recommendation rate, and brand description accuracy.
Half of my clients suffer from inaccurate brand descriptions when users search their brand names directly on AI platforms, mostly caused by inconsistent brand messaging internally.
Lastly, do not feel anxious or succumb to FOMO over GEO. To master GEO fundamentally, learn the underlying logic of LLM and RAG thoroughly, and refine your Prompt Engineering skills.
I have always regarded the claim of "de-AI flavor for AI-generated content" as a false proposition. When AI writing reads rigidly robotic, the problem lies in poor human prompting rather than the AI tool itself. I used AI to write three full books and three million-word novels, and almost no readers can detect AI generation quality. The key lies in your input knowledge and detailed prompts.
The essence of the GEO era is not a new marketing trick, but systematically converting a company’s accumulated professional expertise into digital brand authority. Learning GEO means taking on the role of showcasing your company’s strengths systematically and confidently.
Build strong branding, superior products and high-quality content first. When these foundations are solid, GEO success will follow naturally as an inevitable outcome.
Monica
Fantastic sharing packed with actionable insights. Thank you so much, Uncle Zou. We will organize a book sharing session once the new book is fully released.
Uncle Zou
Thank you all. Looking forward to our next exchange.
No comments yet.