How to Become Visible in AI Search: What Google's Public Liaison for Search Told Us and What He Left Out

Blas Giffuni
4 hours ago
9 min read

Summary:

Looking to become visible in AI? Google's Danny Sullivan confirmed what smart marketers already knew: generic "Top 10 Tips" content is dead in AI search. What works now is content built on real expertise, real clients, and real experience that only you can create.

Good SEO and good GEO are the same thing, but Sullivan missed two things: training data bias creates blind spots for non-English markets, and websites now need to serve AI agents as users, not just people. I tested this with a SaaS client. Five pieces of expert-driven content. Five weeks. Zero to generating leads.

Danny's framework: content must be unique, specific, and authentic. If someone who's never done your job could write it, it's commodity content.
Each LLM surfaces content differently. Visibility in Google doesn't guarantee visibility in ChatGPT or Perplexity.
Your next move: audit your content, start documenting real client stories, and build a specialized content agent to scale your expertise without diluting it.

Danny Sullivan, Google's Public Liaison for Search, laid out the playbook for AI search. Here's what it means for your business and what he missed.

Google's been telling us the same thing for decades: stop creating content for machines and start creating it for people. Most businesses ignored it. They chased keyword density, backlink schemes, and content volume instead.

Now, Danny Sullivan stood on stage and said it again. Only this time, AI search makes it impossible to ignore.

Who is Danny Sullivan? Sullivan serves as Google's Public Liaison for Search. His job is to help the public understand how Google Search works and to collect feedback from the outside community on how search can improve. When he speaks, he's not speculating. He's explaining what Google is doing right now.

What he said at a recent Google event wasn't new. It was a confirmation of a concept I've been presenting to clients and at conferences for years: the user is the algorithm.

The difference today? The algorithm is finally smart enough to enforce it.

The One Sentence That Should Change Your Content Strategy

Danny put a single slide on the screen:

"And good SEO is largely having great content for people."

Not keyword density. Not backlink profiles. Not meta descriptions. Not even structured data, though that, somehow, matters. The foundation of ranking well in AI search is creating content that genuinely serves the person reading it.

If your content strategy is built around gaming an algorithm rather than helping a human, you're building on sand.

What "AI Search" Actually Means for Your Visibility

Danny explained how Google's AI Overview feature works through three layers operating together:

General Knowledge (the AI model itself)
The AI has broad, pattern-based knowledge absorbed from enormous amounts of content. This is how it can answer "Why did the chicken cross the road?" without searching anything. It just knows. You can't influence this layer directly. It's baked into the model.
Specific Knowledge from Traditional Results
For anything current, specific, or nuanced, the AI pulls from real web content surfaced through traditional search ranking. Your website can still be a source, but only if it ranks in the first place.
Fan-Out Queries
This is the part most people miss. When someone searches "red ebikes for a 5-mile commute with hills," the AI doesn't search that exact phrase. It fans out to related queries like "best ebikes," "ebikes for hills," and "red ebikes," then synthesizes the results into a single answer.

You don't need to optimize for one perfect keyword. You need to be the most credible, specific, and authentic source on your topic so you show up across the entire constellation of related queries the AI pulls from.

The Commodity Trap: Why Most Business Content Is Already Dead

A framework introduced during the presentation should make every business and marketing leader uncomfortable: the distinction between commodity and non-commodity content.

Commodity content is the stuff that's everywhere. Generic, recyclable, and increasingly something an AI can write better than you, faster than you, for free. If your content strategy runs on commodity pieces, you're in a race you cannot win.

Here are the examples showed on stage:

Industry	❌ Commodity	✅ Non-Commodity
Running Store	"Top 10 Things to Consider When Buying Running Shoes"	"Why This Customer's Shoes Collapsed After 400 Miles: A Wear Pattern Analysis"
Real Estate Agent	"7 Tips for First-Time Homebuyers"	"Why We Waived the Inspection (And Saved $15k): A Look Inside the Sewer Line"
Interior Designer	"2024 Kitchen Trends You Need to See"	"Marble vs. Grape Juice: Why I Refused to Install Stone for a Family of Five"

The commodity column is advice that any AI, any blogger, any competitor can produce. The non-commodity column is something only you can produce because it comes from a specific experience, a real client, a moment where your expertise showed up in the real world.

The shoe analysis requires someone who actually examined that shoe after 400 miles of wear. The sewer line story exists because the agent crawled the line themselves. The grape juice test happened because the designer ran a real experiment on real marble. No AI can fabricate these. No competitor can copy them.

The Same Two Objections Every Week

When I talk to business owners about content strategy, the pushback is predictable.

The first: "I don't have time or resources for content." This assumes content is a volume game. It's not anymore.
The second: "Can't we just get ChatGPT to produce content in bulk?" This wants to automate the one thing that can't be automated: your expertise. Bulk AI content is commodity content by definition.

Both represent the same misunderstanding. This is the moment for companies to highlight their knowledge, their unique value, and their point of view to differentiate from competitors. It's hard work. I won't pretend otherwise. But the results are there.

Proof: Quality Over Volume Works to Become Visible in AI

We recently worked with a SaaS platform in the go-to-market category. They were invisible in search. Not ranking for anything meaningful.

We didn't flood the internet with content. We created five pieces built entirely around their specific business expertise: real use cases, real insights, real perspectives that only their platform could provide.

Within five weeks, they went from zero visibility to generating qualified leads.

Five pieces. Five weeks. That's what non-commodity content does when you get it right.

What Good Non-Commodity Content Looks Like

Unique. It brings a viewpoint, information, or perspective that others lack or can't easily replicate. Not just a different angle on the same facts. Genuinely different content that couldn't exist without you specifically creating it. Ask yourself: "Could someone who's never done my job write this?"
Specific. It talks about a specific instance, situation, or thing. Not general rules. Not "here are five tips." A real example, a real case, a real result. The more granular you get, the more valuable and the harder it is to replicate. Ask yourself: "Does this reference a real case, client, or result?"
Authentic. It demonstrates first-hand knowledge or experience. You were there. You did the thing. You saw what happened. This is the signal AI cannot fake and that Google is increasingly rewarding. Ask yourself: "Was I there? Did I do the thing?"

If your content doesn't pass all three tests, it's likely commodity.

Good SEO Is Good GEO, With a Caveat

Good SEO is good "GEO" Danny Sullivan, Google

Sullivan addressed the explosion of new acronyms around AI optimization: GEO (Generative Engine Optimization), AEO (Answer Engine Optimization), LLM SEO, AI SEO. He even joked about "LLMNOPEO"; in my case, I created AIAIO.

His point: good SEO is good GEO. The content that ranks well in traditional search is the same content that gets cited in AI Overviews.

I agree, but only to a point.

As far as information retrieval goes, good SEO fundamentals are the foundation of GEO. But Sullivan didn't address training data bias, and this is where the nuance matters.

Bias's Analysis	The Problem	Who It Affects
Language imbalance	English-language content is far more abundant in training data. LLMs often translate from English sources rather than drawing from native-language expertise.	Any business operating in non-English markets
Platform bias	Google and several LLMs lean heavily on Reddit and LinkedIn as a content source.	Non-English-speaking countries and communities, like U.S. Hispanics, where Reddit is largely irrelevant, and LinkedIn is in its infancy.
Model differentiation	Each LLM analyzes content differently to fit its purpose and stand apart from competitors.	Every business, because visibility in one LLM doesn't guarantee visibility in another

I see these dynamics across all major LLMs, not just Google's. ChatGPT search, Perplexity, and others are all surfacing content in their own ways. The principle of quality, authentic content holds across all of them, but the mechanics differ.

If you operate in multilingual or non-English markets, "good SEO = good GEO" needs more nuance than Sullivan suggested.

The Practical Playbook: What to Do Right Now

Danny closed with a list of tips that apply today:

Follow SEO fundamentals. Technical health, page speed, mobile usability, internal linking. None of this goes away.
Use structured data. Schema markup helps Google understand your content and increases the chance of being pulled into AI answers.
Prioritize page experience. A fast, clean, easy-to-use page signals quality. A slow, cluttered, pop-up-heavy page signals the opposite.
Create unique, authentic, non-commodity content. Sullivan called this the single most important factor above everything else.

These align with traditional SEO success. This isn't a pivot. It's an acceleration of what good content marketing was always supposed to be.

What Sullivan Left Out: AI Agents Are the New Users

Here's what I'd add to Sullivan's framework.

Sullivan focused on content being for people. He's right. Content should always serve human readers first.

But the website experience now needs to consider a new type of visitor: AI agents.

AI agents are already browsing websites, extracting information, and making decisions on behalf of users. They're booking appointments, comparing products, pulling data, and summarizing services. This isn't theoretical. It's happening now.

The new rule: Your content is for people. Your website experience needs to work for both people and AI agents.

That means thinking about how your site structure, data formats, and content organization serve not just a human scanning a page, but an AI agent parsing it for answers on someone's behalf.

Sullivan didn't cover this. But it's the next evolution of what "user experience" means in AI search.

The Action Items

Audit your existing content. Go through your last 20 blog posts or web pages. Ask one question: "Could someone who's never done my job have written this?" If yes, that's commodity content. Flag it for revision.
Start a case study habit. Every interesting client situation, every unusual problem you solved, every moment where your expertise showed up: write it down. Even a short internal note is a starting point for non-commodity content.
Make your specificity visible.

❌ Commodity Language	✅ Non-Commodity Language
"We have 15 years of experience"	"Here's what I learned after 15 years of seeing homeowners make the same three mistakes"
"We offer customized solutions"	"Here's how we rebuilt a client's pipeline after their CRM migration failed mid-quarter"
"We're experts in our field"	"This is the test I run before recommending any vendor to my clients"

Think in stories, not tips. Lists of tips are commodity. A specific story with a specific outcome is non-commodity. Every piece of content should have a real protagonist: a client, a situation, a decision. Not abstract advice.
Document your reasoning, not just your conclusions. The interior designer didn't just say "don't use marble with kids." She showed the grape juice test. The real estate agent didn't just say "waive the inspection." He explained why, because he crawled the line himself. Your reasoning process is as valuable as your answer.
Build your own specialized content agent. I built mine. Her name is Alicia. She helps me write every piece of content you see from Navigamo. But here's the key: Alicia won't produce anything unless I provide the guidance, the expertise, and the perspective first. She's a multiplier of my knowledge, not a replacement for it.
That's the right way to use AI for content. The agent handles structure, editing, and consistency. You provide the non-commodity ingredient: your expertise.
If you want to build your own content agent, I recommend nemo® as a platform. It's easy to use and purpose-built for this. I wrote about how I use AI agents to multiply content output without sacrificing quality if you want to see the process in detail.
Invest in video content and give it a proper home. Video is becoming even more critical in AI search. LLMs are increasingly pulling from video content, and Google continues to prioritize video in results. But uploading to YouTube alone isn't enough. Create a dedicated landing page on your website for each video with a full transcript, supporting written content that adds context, and structured data (VideoObject schema). A video without a proper landing page is a missed opportunity for both traditional and AI search visibility.

Final Thought

The commodity era of content is ending. The specificity era is here.

Danny Sullivan didn't reveal a secret. He confirmed something that thoughtful marketers have known for years: the best content wins. What's changed is that the algorithm now has enough intelligence to tell the difference.

The question is whether your business will adapt or keep publishing "Top 10 Tips" articles into the void.

Based on a presentation by Danny Sullivan, Google's Public Liaison for Search, on AI in Google Search.