Let’s start with a bit of honesty:
No one is an expert in this space—not yet, and definitely not me.
Whether we’re calling it AEO (Answer Engine Optimization), LLMO (Large Language Model Optimization), AIO (AI Optimization), or any of the new acronyms people are still inventing—this is a completely new frontier. Things are shifting fast, and much of what’s true today might evolve tomorrow.
But one thing is already clear:
AI is changing how people search, discover, and decide.
And that raises a big question for brands:
How do we make sure we show up in AI-generated answers?
From global platforms like ChatGPT, Gemini, and Perplexity to Chinese systems like DeepSeek, Doubao, Yuanbao, and Kimi—the ways people interact with information are being rewritten. And we, as marketers, SEO specialists, and brand builders, need to start thinking differently too.
A Quick Shoutout Before We Dive In
Aleyda Solis wrote a fantastic piece recently:
“SEO vs GEO: Optimizing for Traditional vs. AI Search”
It’s a smart overview for anyone exploring this space and well worth your time.
Some of what we’ll discuss here overlaps, but our goal is to go a little deeper into the hands-on tactics and strategic thinking we’ve been working on—especially from our perspective at Jademond, where we support both international SEO and Chinese digital marketing every day.
I’m not pretending to have all the answers.
But I do have a few thoughts on how brands can position themselves in this new world—and I’d love to hear yours too.
👉 Join the conversation on https://www.linkedin.com/posts/marcuspentzek_are-you-already-optimizing-your-brand-for-activity-7335945365193428993-bK13 — this is an ongoing learning journey, and your perspective matters.
Let’s dig in.
Table of Contents
Quick Note on Search vs. AI Use
People use traditional search engines—like Google, Baidu, or Bing—differently than they use what we loosely call “AI” tools (though let’s be honest, they’re not intelligent, just very good at predicting language). Aleyda’s article and the Google Spreadsheet accompanying her post does a really good job in pointing out the differences, so we skip it here.
That said, users are turning to these LLM-based tools—like ChatGPT, Gemini, DeepSeek, Doubao, and others—to ask questions and do research. In both cases, whether through a search engine or an AI assistant, the outcome is similar: brand names may appear either as sources of information or as recommended providers.
And that’s the opportunity we’re focused on.
This article is all about how to position your brand to show up in those answers—whether you’re being cited, referenced, or recommended.
Where AI Answers Come From: The Role of Training Data
Before we talk tactics, it’s important to understand one thing: LLMs (Large Language Models)—the engines behind tools like ChatGPT, Gemini, DeepSeek, Kimi, and others—don’t “know” things the way humans do. They’re not intelligent. They’re just trained to predict what comes next in a sentence based on patterns they’ve seen in huge amounts of text.
And where does that text come from?
A big chunk of it comes from publicly available, online-accessible content:
- Open source knowledge bases like Wikipedia or Baidu Baike
- User-generated content platforms like Reddit, X (Twitter), Weibo, Zhihu
- Forum-style discussions and publisher networks like Baidu Baijiahao
During training, these models break down the language they see into patterns, learning which words and phrases tend to appear together, and in what context. Over time, the model builds up probabilities—essentially: “When people talk about brain health, what words or ideas usually show up?”
Here’s a (very) simplified example:
If the majority of nutrition articles in the training data mentioned that “omega-3 fatty acids support brain function,” and that message appeared across trusted sources, then that’s the kind of answer the AI is likely to generate when asked how to support brain health.
If—hypothetically—all those sources instead said “Coca-Cola boosts brain fitness,” then the model would probably parrot that back too. (Thankfully, it doesn’t.)
The key idea:
LLMs don’t know what’s true—they reflect the patterns of what they’ve seen.
So if your brand or product is consistently mentioned in helpful, topic-relevant content online, that increases the chance it will show up in the AI’s answers. That’s the game.
And that’s what the rest of this article is about: how to shape those patterns in your favor.
Worried Your Brand Wasn’t in the Training Data? You’re Not Alone.
If you’re a smaller or newer brand, you might be thinking:
“Well, that’s it—we weren’t part of the training data, so we’re out of the game before it even started.”
That fear is understandable—but here’s the good news:
The training isn’t over.
While most major LLMs (like GPT-4, Gemini, or those powering DeepSeek and Kimi) were initially trained on massive datasets from the past, they don’t stay frozen in time. The companies behind these models are constantly working to update and refine them, so their outputs don’t become outdated or irrelevant.
This happens in a few ways:
- Continual data ingestion:
Many LLM providers feed in fresh online content—from news sites, forums, databases, and community platforms—to help keep the model aware of current events, new products, trends, and brand mentions. - Partial re-training & pattern adjustment:
Even without rebuilding the entire model, engineers can fine-tune parts of the model or adjust weighting by training on new corpora of text. This helps shift the model’s understanding and associations over time—yes, even to include newer brands and ideas. - Prompt fine-tuning (instruction tuning):
Beyond knowledge, models are constantly being improved on how they respond—whether that’s recommending products, comparing options, or citing sources more clearly.
What this means for you:
Even if your brand wasn’t part of the original training data—you still have a shot.
By creating content that is findable, helpful, and consistently associated with the topics you want to “own,” you can begin to show up in the places that matter: both in traditional search and in AI-generated answers.
The Second Path: Real-Time Lookups & Live Data
There’s another important way large language models bring in information—they don’t always rely only on what they were trained on.
Depending on the tool, the prompt, and the user’s settings, some AI assistants perform live research. This might mean:
- Querying their own search index, if the AI provider maintains one (though few do).
- Tapping into traditional search engines like Bing, Google, or Baidu.
- In China, using platform-specific indexes, often built from social media platforms, news outlets, or even crawled websites, depending on the provider (e.g., Baidu’s own index, or ByteDance and Tencent’s internal data sources).
This kind of real-time retrieval allows the model to pull up-to-date information and even cite sources in some cases—especially for time-sensitive queries or questions about emerging topics and brands.
This is where traditional SEO still matters—big time.
If your website, product pages, or brand content is well-optimized and findable through search, there’s a higher chance those AI systems will pick up your content during live lookups and include it in their responses.
So even as we talk about “AI Optimization,” don’t forget:
A solid SEO foundation is still part of the game. It makes your content available not just to humans—but to the machines doing the summarizing and recommending on their behalf.
How to Position Your Brand in AI Answers: A Practical Strategy
Now that we understand how LLMs get their information, it’s clear that visibility in AI-generated answers depends on three things:
- Your content being findable
- Your brand being mentioned by trusted sources
- And ideally, being mentioned by many different sources, not just one
So, how do we start influencing that?
Here’s a simple but powerful 3-part strategy:
1. Traditional SEO – Still Foundational, but Smarter
Yes, we’re talking about Answer Engine Optimization, but don’t let the shiny new acronym fool you: traditional SEO is still your foundation.
AI tools like Gemini and Microsoft Copilot often lean on their own crawling and indexing capabilities, or on existing search engines like Google and Bing—which means your website content needs to be both findable and understandable to those systems.
Here’s how to approach it with AI visibility in mind:
✅ Cover the Full Range of Questions People Might Ask
When someone asks an AI a question that’s related to your business, product, or industry, you want your content to be part of the answer.
So start by building a solid content strategy that addresses:
- Questions directly about your brand (Who are you? What do you do? Why should people trust you?)
- Questions about your product or service category (e.g., What’s the best type of air purifier for small apartments?)
- Problem-based queries (e.g., How do I reduce indoor allergies?)
Think like a user. Think like a customer. Think like an LLM prompt.
Then build high-quality, helpful content that answers those questions in natural, clear language.
✅ Use Your Brand Name in the Content
Here’s a subtle but powerful shift to make:
Most writers don’t mention the brand name much on their own website—after all, the user is already there, right?
But LLMs don’t have that context. They don’t know what site they’re reading from. They just process text.
So if your website says, “Our team believes in clean skincare,” the model sees a belief… but not who holds it. If instead you say, “Our team at Jademond believes in clean skincare,” then your brand becomes attached to that belief.
Repeating your brand name (naturally) throughout your website helps models form stronger associative patterns between your brand and your topic.
Don’t overdo it. But don’t be shy either.
✅ Add Structured Data (Schema.org)
Structured data—especially Schema.org markup—helps machines understand your content on a deeper level.
Google has made it clear for years that structured data helps its bots better interpret and classify content. And since Gemini is built on top of Google Search, that’s a strong signal it’s in use there. Similarly, Microsoft Copilot is confirmed to use structured data in tandem with Bing results to inform its responses.
So make sure you’re using structured data to mark up:
- Articles & blog posts (Article, NewsArticle, BlogPosting)
- Products (Product, Offer, Review)
- FAQs (FAQPage)
- How-tos (HowTo)
- Your business (Organization, LocalBusiness)
Even if a chatbot doesn’t directly cite your site, this markup gives it better odds of understanding what your brand does—and why it matters in a given context.
But: Structured data does only assist – it is not a magic tool. Make sure to focus on the good content first!
The bottom line:
Traditional SEO—especially with thoughtful content and structured data—remains the bedrock of AI visibility. Don’t think of it as outdated. Think of it as step one in making your brand machine-readable and AI-ready.
2. Digital PR – Earn Mentions That Actually Matter
Let’s face it—most digital PR is forgettable.
It’s press releases no one reads, generic guest posts, and “me too” announcements trying to look important.
But in the world of Answer Engine Optimization, that kind of content won’t help much.
Here’s the truth:
LLMs pick up on patterns across many sources—especially those that are trusted, well-read, and referenced often.
So you don’t just want to be mentioned. You want to be talked about in context, and in places that machines (and people) trust.
Here’s how to approach it with AI visibility in mind:
✅ Shift from Self-Promotion to Real Value
Forget writing about your new office or your award unless it’s genuinely interesting. Instead, focus on stories that:
- Solve a common problem in your industry
- Offer original data or insights
- Share a bold opinion or trend analysis
- Connect emotionally or culturally with your target audience
Create content that someone would actually want to:
- Bookmark
- Share on LinkedIn
- Refer to in a meeting
- Use in their own writing
If your brand is part of a real answer, it has a much better chance of showing up in AI-generated answers.
✅ Mentions Matter More Than Links
In traditional SEO, we obsessed over backlinks. But for LLM visibility, brand mentions—especially in natural-language content—are just as (if not more) important.
So focus on:
- Getting your brand named in the body of articles, ideally in ways that associate you with your field of expertise
- Being quoted as a source
- Publishing or contributing to content that aligns with topics you want to be surfaced for in AI answers
Example: If you’re a B2B SaaS brand helping with logistics optimization, aim to be mentioned in content discussing “how AI is improving logistics efficiency”—not just “10 startups to watch.”
✅ Go Beyond “We Wrote This” Mentions
Publishing content yourself on other platforms is a good start, but stronger signals come when others mention your brand independently.
So yes, by all means:
- Publish thought leadership on Medium, Zhihu, LinkedIn Articles
- Get featured on relevant blogs, WeChat Official Accounts, and industry newsletters
But also:
- Pitch ideas to journalists and editors
- Partner with influencers and experts who actually write things people trust
- Create things people naturally talk about, not just read once
And when you do have a hand in getting mentioned (say, through a contributor piece or PR push), don’t just wedge your brand in. Make it:
- Memorable
- Relevant
- Easy to associate with your core expertise
The goal isn’t just coverage—it’s context.
If your brand is mentioned where the right conversations are happening, AI tools will start to associate you with those topics.
Authority Mentions – Be Where the Machines Trust
If Digital PR helps you get talked about, Authority Mentions help you get remembered.
LLMs—just like people—tend to trust certain sources more than others. Whether it’s because they appear frequently in training data, have high domain authority, or consistently produce quality content, these trusted platforms act as signal boosters for your brand.
In short: If you’re mentioned by a source the model already respects, your brand is more likely to be included in its answers.
Here’s how to approach it with AI visibility in mind:
✅ Understand the Role of “Trusted Sources” in LLMs
We know that models like ChatGPT, Gemini, and Copilot have been trained on vast corpora of online content—including high-authority domains like:
- News media (e.g. The Guardian, Forbes, TechCrunch)
- Wikipedia, Baidu Baike
- Government and educational institutions
- UGC platforms with strong moderation (e.g. Reddit, Zhihu, StackExchange)
- Major Q&A and editorial platforms (e.g. Quora, Zhidao, Zhihu)
Even in China, engines like Doubao or Kimi are informed by what people read, share, and cite across WeChat, Baijiahao, Xiaohongshu, and so on.
So if your brand appears in content published by or referenced in those environments, it earns more attention—not just from people, but from language models too.
✅ Don’t Just Pay for Placement—Create Something Worth Citing
Yes, there are ways to get into top-tier publications. Agencies know it. Journalists know it. The world knows it.
But when that happens, make it count.
Don’t waste the opportunity on:
- A fluffy announcement
- A generic product plug
- An “interview” no one really reads
Instead:
- Publish something data-driven, surprising, or contrarian
- Make sure your brand is clearly tied to a topic or insight
- Include a quote or claim that’s quotable—something that others might reference later
If you have the access, earn that mention with substance.
✅ Play the Long Game: Relevance Beats Hype
Smaller brands often feel discouraged here—after all, not everyone gets into Forbes.
But there are plenty of micro-authoritative spaces where your brand can shine:
- Niche industry blogs and media outlets
- Academic or white paper collaborations
- Influencer or KOL content in specific regions or platforms
- Localized trade publications (e.g., “SaaS Weekly China” or “Ecom Watch Europe”)
Find the ones trusted in your field, and consistently offer value that helps you stay visible—not just in one article, but as a reliable voice over time.
The endgame isn’t just prestige—it’s position.
You want your brand to become part of the LLM’s internal map of “who knows what.”
And Authority Mentions? They light up the map.
User-Generated Content (UGC) & Community Signals – What People Say, Models Repeat
You can have great SEO. You can get mentioned in the press.
But if real people aren’t talking about your brand, your visibility in LLM-generated answers might still fall short.
Why?
Because LLMs are trained to sound like people. And that means they learn from the places where people talk.
That includes:
- Reddit threads and Quora answers
- Xiaohongshu posts and Weibo discussions
- Product reviews and Amazon Q&A
- Forum posts, niche Discords, Telegram groups
- Blog comments and YouTube discussions
- Zhihu answers or Baidu Tieba threads
It’s the noisy, messy, very human part of the internet—and it heavily shapes how models interpret what’s credible, liked, and repeated.
Here’s how to approach it with AI visibility in mind:
✅ Encourage Organic, Real-World Brand Mentions
You can’t fake community. But you can encourage it.
Think about ways to:
- Get customers talking about your product in authentic ways (prompted reviews, challenges, contests)
- Seed useful commentary or advice under your brand name on community platforms
- Collaborate with micro-influencers or KOLs who are deeply embedded in real conversations
- Ask users to share their experience in the exact forums where people seek advice
Example: If people ask Gemini, “What’s a good CRM for small startups in Asia?”—you want your brand to have been named in Reddit threads, Zhihu answers, or product comparison blog posts that were likely ingested or indexed.
✅ Be Present Where the Prompts Happen
Users don’t always go to Google anymore. They go to:
- X/Twitter to ask their network
- Reddit to get real, unfiltered opinions
- Xiaohongshu to see how something feels or works
- Zhihu to get reasoned, long-form explanations
- Weibo, Baijiahao or niche Douban groups to hear what friends and experts are saying
These are places where prompt-like behavior happens daily—and they shape the expectations and style of AI-generated answers.
So ask yourself:
- Is my brand present in those spaces?
- Am I encouraging discussions around the problems we solve?
- Do I engage where it makes sense, not just broadcast?
✅ The LLMs Are Watching (and Remembering)
Some LLMs were trained on static datasets from before 2023. Others, like those powering Gemini or DeepSeek, are being updated continuously or through retrieval mechanisms. That means they are:
- Monitoring real-time forums
- Ingesting scraped public content
- Learning from platforms with high engagement and trust
The more people talk about your brand naturally—in helpful, positive, or insightful ways—the more likely LLMs will remember those patterns and reflect them in answers.
What matters to people, matters to AI.
So build a brand people want to talk about—and do it in the places where real conversations shape machine learning.
Wrapping It Up: No Experts, No Snake Oil—Just Smart Experimentation
Let’s be clear about one thing:
Nobody has all the answers when it comes to AEO, LLMO, AIO—or whatever acronym we’re using next.
There are no true experts in this space yet. Not because people aren’t smart, but because:
- The technology is changing rapidly
- Each AI system (ChatGPT, Gemini, Kimi, Doubao, Deepseek, etc.) is built differently
- Their training sources, update cycles, and answer mechanisms vary wildly
- Even the companies behind them are still figuring things out
So if someone tells you they’ve “cracked the AI ranking algorithm”?
They’re either selling snake oil, or they don’t understand the complexity they’re dealing with.
What we can do is apply real digital marketing principles—strong content, brand visibility, technical clarity, and strategic positioning—into this new environment where machines help shape answers.
And we can do it with humility, curiosity, and iteration.
This guide is just a starting point. I don’t claim to have it all figured out. But these are the ideas and patterns we at Jademond have observed, tested, and believe are worth exploring.
💬 I’d love to hear your thoughts, too—
What’s working for you? What are you experimenting with?
Let’s keep the conversation going. I’ve opened a thread on LinkedIn here: https://www.linkedin.com/posts/marcuspentzek_are-you-already-optimizing-your-brand-for-activity-7335945365193428993-bK13 — jump in and share!
Together, we’ll figure out how to help our brands earn a place in the answers—wherever, and however, people are asking.
About the author: Marcus Pentzek, Director of SEO, Jademond Digital - Marcus Pentzek has been shaping the SEO landscape since 2008, beginning as a consultant in Germany and later pioneering SEO strategies at Searchmetrics GmbH. His deep dive into the Chinese market began in 2012 while directing marketing at Yoybuy Ltd in Beijing, gaining firsthand experience in e-commerce SEO in China. Since 2022, he leads SEO at Jademond Digital, focusing on innovative, data-driven methods tailored for Chinese audiences. His blog posts merge over a decade of global SEO expertise with practical insights into the Chinese digital environment.