Introduction: GEO Is More Than Getting Mentioned by ChatGPT

Generative engine optimization (GEO) is the practice of improving your content and digital presence so your information has a better chance of being discovered, understood, referenced, and surfaced in AI-powered search experiences.

But there’s a problem with how GEO is discussed online.

A lot of articles make it sound like there’s a secret formula:

Add more entities.

Create llms.txt.

Write short paragraphs.

Add schema.

Mention ChatGPT.

Get backlinks.

Then wait for AI to cite you.

That’s not how I’d approach it.

The reality is more interesting.

AI-powered search is changing the way information is retrieved and presented, but the fundamentals behind discoverability haven’t disappeared. Google’s current guidance says its AI Overviews and AI Mode continue to rely on core Search systems, including retrieval from Google’s Search index.

At the same time, academic research has shown that content optimization can influence visibility in generative search systems.

The 2023 research paper “GEO: Generative Engine Optimization“, authored by researchers including Pranjal Aggarwal, Vishvak Murahari, Tanmay Rajpurohit, Ashwin Kalyan, Karthik Narasimhan, and Ameet Deshpande, introduced GEO as a framework for improving content visibility in generative-engine responses. In its experiments, the researchers reported visibility improvements of up to 40%, although the effectiveness of different tactics varied across domains.

That gives us a much better starting point.

GEO isn’t about writing for a machine.

It’s about creating information that is:

• Discoverable
• Understandable
• Useful
• Specific
• Well-supported
• Contextually relevant
• Easy to connect with related information
• Worth referencing

And increasingly, that information isn’t limited to text.

AI search can surface web pages, images, videos, products, local information, and other content formats. Google explicitly recommends supporting textual content with high-quality images and videos where appropriate for its generative AI experiences.

So if you’re still thinking about GEO as “SEO for ChatGPT,” you’re thinking too narrowly.

GEO is really about becoming a better source of information in an increasingly generative search environment.

What Is Generative Engine Optimization?

Generative Engine Optimization is the process of improving content, website structure, entities, authority, and supporting media so information can be more effectively discovered and used within generative search experiences.

The term became prominent after the 2023 research paper GEO: Generative Engine Optimization. The researchers described generative engines as systems that synthesize information from multiple sources and generate responses rather than simply presenting traditional search results.

That distinction matters.

Traditional search might look like:

Query → Search results → User chooses a page

A generative search experience can look more like:

Question → Retrieval → Multiple sources → Synthesis → Generated response → Supporting citations

The exact implementation differs by platform.

But the user’s experience is fundamentally different.

Instead of deciding which ten results to open, the user may receive a synthesized answer first and then decide which sources deserve further investigation.

That creates a new visibility opportunity.

Your website doesn’t only need to rank.

Your information needs to be useful enough to become part of the answer.

GEO vs SEO: What’s the Difference?

GEO is not a replacement for SEO.

This is one of the most important points in this entire guide.

Google’s current documentation is explicit: its generative AI search features are built on core Search ranking and quality systems, and the same foundational SEO practices continue to matter.

Here’s a simpler comparison:

Area	Traditional SEO	Generative Engine Optimization
Primary goal	Organic search visibility	Visibility within AI-generated search experiences
Core asset	Web page	Information source
Search behavior	Keywords and queries	Questions, context and multi-step intent
Discovery	Search results	Search + AI retrieval
Visibility	Rankings	Citations, mentions, references and AI impressions
Content	Search-optimized pages	Clear, useful, evidence-backed information
Entities	Important	Increasingly important for context
Structured data	Useful for search features	Useful as part of broader machine-readable SEO, but not a special GEO requirement
Images/video	Supporting assets	Can become additional sources of AI-search visibility
Measurement	Rankings, clicks, traffic	AI visibility + traditional SEO + business outcomes

The relationship is better represented like this:

SEO is the foundation. GEO is the additional layer of thinking required when search becomes generative.

That means your GEO strategy should not begin by throwing away your SEO strategy.

It should begin by asking:

What parts of our existing SEO system need to evolve for AI-assisted search?

Where the GEO Concept Came From

Before discussing tactics, let’s go back to the research.

The paper “GEO: Generative Engine Optimization” was published as an arXiv preprint in November 2023. It introduced a framework for improving the visibility of content in generative engine responses and created the GEO-bench benchmark for evaluating different optimization strategies.

The researchers tested different approaches to content optimization.

Their findings showed that optimization could improve visibility substantially in their experimental environment, with improvements of up to 40%.

But there’s a critical nuance.

The paper did not establish a universal rule that:

“Add statistics and your citations will increase by 40%.”

That’s an oversimplification.

Different tactics performed differently depending on the domain and evaluation setting.

That nuance is important because GEO is still a rapidly changing field.

The research gives us evidence that optimization can matter.

It does not give us a permanent ranking formula.

What Did the Original GEO Research Actually Teach Us?

One useful lesson from the study is that how information is presented can influence how visible it becomes in generative responses.

The researchers evaluated several content optimization approaches, including techniques involving:

• Citations
• Statistics
• Quotations
• Fluency
• Technical terminology
• Content structure
• Other forms of optimization

The effectiveness varied.

That leads to a better GEO principle:

Don’t optimize for one tactic. Optimize the information itself.

For example, statistics can strengthen an article when they are relevant and credible.

A quotation can add authority when it comes from a legitimate expert.

A citation can improve trust when it points to a primary source.

Technical terminology can help when you’re writing for an expert audience.

But adding any of these elements artificially doesn’t automatically make content better.

Context matters.

How Does AI Search Retrieve Information?

This is where GEO gets more technical.

A simplified model looks like this:

User Question    --> Query Understanding    --> Search / Retrieval    -->  Relevant Sources     -->  Content Evaluation      -->  Answer Generation      -->  Citations / Supporting Links

This isn’t a universal architecture for every AI search product.

Different systems use different models, retrieval systems, ranking layers, indexes, and interfaces.

But Google’s current documentation gives us a useful example.

Google says its AI search features can use retrieval-augmented generation (RAG), where relevant pages are retrieved from Google’s Search index and used to ground generated responses. Google also describes query fan-out, where related queries can be generated to investigate different aspects of a user’s question.

This matters because it changes how we should approach content.

You don’t need to predict every possible question.

You need to create content that provides strong answers to meaningful questions within your topic.

What Is Retrieval-Augmented Generation?

Retrieval-augmented generation, or RAG, combines information retrieval with generative AI.

Instead of relying only on what a model already knows, the system can retrieve external information and use it to generate a response.

A simplified version is:

Question → Retrieve → Ground → Generate

For website owners, the important part is the middle.

Retrieve.

Your content needs to be accessible and discoverable.

This is why traditional technical SEO still matters.

Google says pages need to be indexed and eligible for Search in order to be eligible as supporting links in AI Overviews or AI Mode. There aren’t separate technical requirements just for these AI features.

Does GEO Mean You Need to “Chunk” Your Content?

This is one area where the industry has gone too far.

You’ll often hear advice like:

“Break every article into tiny 50-word chunks because AI systems store everything in vector databases.”

That sounds technical.

It isn’t necessarily accurate.

We don’t have a universal blueprint for how every generative search system stores or retrieves web content.

More importantly, Google explicitly says there is no requirement to break content into tiny pieces for AI search. Google says its systems can understand the nuance of multiple topics on a page and that there is no ideal page length.

So what should do?

Use logical information units.

For example:

What Is GEO?

Give a clear definition.

Why Does GEO Matter?

Explain the problem.

How Does AI Search Work?

Explain retrieval and generation.

How Do You Measure GEO?

Give the metrics.

This isn’t “vector database optimization.”

It’s simply good information architecture.

And that’s the point.

Don’t create artificial chunks for an AI crawler. Create useful sections for humans that are also easy for machines to interpret.

Query Fan-Out: Why Topical Depth Matters

Google’s AI search documentation describes query fan-out as a technique where systems can issue multiple related searches to address different aspects of a user’s question.

Imagine a user asks:

What is the best CRM for a small SaaS company?

A useful answer might require information about:

• CRM pricing
• CRM features
• SaaS sales workflows
• Team size
• Automation
• Integrations
• Reporting
• Ease of implementation

The original query is only one sentence.

The underlying information need is much larger.

This is why topical coverage matters.

Your website should answer related questions because your audience needs those answers, not because you want to create pages for every theoretical fan-out query.

Google specifically warns against creating large numbers of pages primarily to manipulate rankings or generative AI responses.

That’s an important distinction.

Topical depth is good. Programmatic fluff is not.

The New GEO Content Model

Instead of thinking:

Keyword → Article

I’d recommend:

Entity → Topic → Intent → Question → Answer → Evidence → Context → Supporting Media

Let’s break that down.

Entity

What is the article about?

Example:

Generative Engine Optimization

Topic

What aspect are you discussing?

AI citations

Intent

What does the user want?

Understand how citations work

Question

What are they actually asking?

How do AI search engines decide which websites to cite?

Answer

Give the direct explanation.

Evidence

Support it with documentation, research or examples.

Context

Connect it to SEO, content quality, retrieval and authority.

Supporting Media

Add diagrams, screenshots, videos or other relevant visual assets where useful.

That’s a much stronger content model than simply inserting keywords.

The Five Layers of a GEO-Ready Content Asset

I use a five-layer model when thinking about GEO.

Layer 1: Discoverability

Can the content be found?

Check:

• Crawlability
• Indexability
• Internal links
• XML sitemap
• Canonicals
• HTTP status
• Search accessibility

Layer 2: Comprehension

Can the system understand the content?

Use:

• Clear headings
• Direct definitions
• Logical sections
• Descriptive language
• Semantic relationships
• Consistent terminology

Layer 3: Evidence

Why should anyone trust the information?

Use:

• Primary sources
• Research
• Statistics
• Expert quotes
• Original data
• Case studies
• First-hand experience

Layer 4: Context

Where does this information fit?

Connect the content to:

• Related topics
• Entities
• Internal pages
• External references
• Products
• Services
• Authors
• Organizations

Layer 5: Multimodal Value

Can the information be understood through more than text?

Consider:

• Images
• Diagrams
• Videos
• Screenshots
• Charts
• Product visuals
• Demonstrations

Google explicitly recommends high-quality images and videos as part of succeeding in its generative AI search experiences.

The Technical Side of GEO

This is where the original article was too generic.

Technical SEO isn’t just:

Check your sitemap.

For GEO, the more useful question is:

Can search and AI systems access the actual information that matters on the page?

That changes the audit.

1. Make Important Information Available as Text

Google’s AI-search guidance specifically recommends making important content available in textual form.

This matters because some websites hide critical information inside:

• Images
• Canvas elements
• Video-only content
• Interactive widgets
• Client-side JavaScript
• PDFs
• Visual calculators

That doesn’t mean these formats are bad.

It means you shouldn’t make them the only place where important information exists.

For example:

Don’t put your entire product comparison inside an infographic.

Put the core comparison in HTML text too.

Then use the infographic as a visual enhancement.

2. Use Semantic HTML—But Don’t Obsess Over It

Semantic HTML is useful.

Use:

• <article>
• <section>
• <header>
• <nav>
• <main>
• <h1>
• <h2>
• <h3>
• <table>
• <ul>
• <ol>

But don’t turn this into another GEO myth.

Google’s current documentation says perfect semantic HTML isn’t required for its AI search systems. However, semantic HTML can help other users, including people using screen readers, navigate pages.

So use semantic HTML because it’s good web development.

Not because someone promised it would make an AI cite you.

3. Don’t Hide Your Primary Answer

This is one of the most practical GEO recommendations.

If your page answers:

What is GEO?

Don’t make users scroll through 800 words before finding the definition.

Use a structure such as:

Generative Engine Optimization is…

Then explain.

Then expand.

This creates a clear information hierarchy.

4. Build Logical Content Sections

A section should answer a meaningful question.

For example:

H1: Generative Engine Optimization
H2: What Is GEO?H2: How Does AI Search Work?
H2: What Makes Content Citation-Worthy?
H2: How Does RAG Affect Content Discovery?
H2: How Should You Structure Content?
H2: How Do Images and Videos Affect GEO?
H2: How Do You Measure AI Visibility?
H2: Common GEO Mistakes
H2: GEO Checklist

That’s useful information architecture.

You don’t need to split every paragraph into an individual “AI chunk.”

5. Internal Linking Becomes More Important

Internal links establish relationships between pages.

For a GEO content cluster, think:

Pillar

↓

Supporting articles

↓

Sibling articles

↓

Supporting resources

Structured Data: What It Can and Cannot Do

Structured data is often overhyped in GEO discussions.

Let’s separate fact from speculation.

What schema can do

Structured data can help search engines understand specific information about a page and can support eligibility for certain Search features and rich results.

Examples include:

• Article
• Organization
• Person
• Product
• LocalBusiness
• BreadcrumbList
• VideoObject

What schema cannot guarantee

Schema does not guarantee:

• AI citations
• AI Overview inclusion
• ChatGPT mentions
• Perplexity citations
• Gemini visibility
• Higher rankings

Google’s current guidance explicitly says there is no special Schema.org markup required for generative AI search.

So schema belongs in your technical SEO strategy.

Not in a fantasy “AI citation hack.”

Entity SEO and GEO

Entities become particularly useful when you stop thinking only in keywords.

Imagine a company called Conversion Boost.

The website might contain:

Organization: Conversion Boost

Founder: [Person]

Service: Conversion Rate Optimization

Service: Website Conversion Audit

Topic: CRO

Topic: A/B Testing

Topic: Landing Page Optimization

Article: How to Find Conversion Leaks

Case Study: SaaS Conversion Improvement

Those relationships provide context.

The goal isn’t to manipulate an AI model.

The goal is to make the business easier to understand.

That means:

• Consistent brand name
• Consistent organization information
• Clear author profiles
• Descriptive service pages
• Related content
• Internal links
• Appropriate structured data
• External references

Evidence Is a GEO Asset

This is one area where the original article needed much more depth.

If you’re writing about:

How much does SEO cost?

Give actual ranges only when you have reliable evidence.

If you’re writing:

Does CRO improve conversions?

Don’t simply say:

“CRO can dramatically increase conversions.”

Show the evidence.

If you’re making a claim about AI search:

Cite the platform’s documentation.

If you’re discussing GEO research:

Cite the original paper.

If you’re discussing a Google feature:

Use Google Search Central.

This creates a stronger information asset.

Statistics: Use Them Correctly

The original GEO research is particularly relevant here.

The 2023 GEO study tested content optimization strategies including statistics and reported that optimization could increase visibility by up to 40% in its experimental setting.

But here’s how I would apply that finding.

Don’t write:

“Add three statistics to every article and AI will cite you.”

Instead:

“Use relevant statistics when they genuinely strengthen the answer, and cite their sources.”

That’s a completely different philosophy.

A statistic should answer:

Why does this matter?

Not:

How can I make this page look authoritative?

Expert Quotes: Another Layer of Evidence

Quotes can add useful context.

For example, if you’re writing about Google’s AI search:

Google says its generative AI experiences continue to rely on core Search systems.

That’s stronger when supported by Google’s documentation.

But don’t manufacture quotes.

Don’t paraphrase a source and put it inside quotation marks.

Don’t quote someone just because the quote contains a keyword.

Use expert commentary when it actually improves the reader’s understanding.

Multimodal GEO: The Part Most SEO Teams Are Missing

This is probably the biggest gap in many GEO discussions.

Search isn’t text-only anymore.

Google explicitly says its generative AI experiences can surface relevant images and videos, and recommends supporting text with high-quality visual content where appropriate.

That creates another optimization layer.

Call it multimodal search visibility.

Image Optimization for GEO

Don’t treat images as decorative elements.

A useful image can provide information.

For example, if you’re writing about:

How to perform a technical SEO audit

instead of using a generic stock image, create:

• An SEO audit workflow
• A crawl diagram
• A screenshot of a technical issue
• A before/after example
• A visual checklist

Google recommends making images discoverable and optimizing the pages on which they appear.

Use:

• Descriptive filenames
• Useful alt text
• Relevant surrounding text
• High-quality images
• Appropriate dimensions
• Proper image indexing practices

Don’t stuff keywords into alt text.

Write it for accessibility first.

Video SEO and GEO

Video is another opportunity.

Suppose you publish:

How to Find Conversion Leaks in a Website

Your article could include a three-minute video showing the audit process.

The page then contains:

Text explanation

Screenshots

Video walkthrough

Checklist

That’s much richer than 2,000 words of text.

For instructional content, video can demonstrate things that are difficult to explain through prose.

Examples:

• SEO audits
• Software tutorials
• Product demonstrations
• Technical fixes
• Data analysis
• Design walkthroughs
• How-to processes

Google’s current AI-search guidance specifically recommends supporting text with high-quality images and video where relevant.

Don’t Forget YouTube

For certain topics, YouTube can become part of the search visibility ecosystem.

Consider creating:

• Explainer videos
• Tutorials
• Case studies
• Product demonstrations
• Expert interviews
• Webinar clips

Then embed relevant videos into the corresponding article.

Don’t create a video simply because “GEO says video is important.”

Create it because the information is easier to understand visually.

Diagrams Are Underrated

A good diagram can compress a complicated concept.

For example:

USER QUESTION                     
↓              
QUERY ANALYSIS                     
↓              
RELATED SEARCHES                     
↓              
SOURCE RETRIEVAL                     
↓          
CONTENT + ENTITY CONTEXT                     
↓              
AI GENERATION                     
↓            
ANSWER + CITATIONS

That’s easier to understand than three paragraphs explaining the same process.

For technical subjects, diagrams can be some of the highest-value assets on the page.

GEO for Product and Ecommerce Websites

GEO isn’t only for blogs.

Product information is particularly suited to structured, factual content.

For example, an ecommerce product page can clearly communicate:

• Product name
• Brand
• Features
• Specifications
• Materials
• Dimensions
• Availability
• Price
• Reviews
• Images
• Videos
• Use cases

Google’s AI-search guidance also highlights shopping and product information as areas where generative experiences can surface different types of content.

For ecommerce brands, GEO should therefore connect:

Product SEO + Structured Data + Merchant information + Images + Video + Reviews + Content

rather than treating the product description as the entire strategy.

GEO for Local Businesses

Local businesses have another advantage.

Their information exists across multiple sources.

Think about:

• Google Business Profile
• Website
• Reviews
• Local directories
• Social profiles
• Local publications
• Maps
• Business listings

Consistency matters.

Your:

Business name

Address

Phone number

Services

Opening hours

Categories

should be accurate and consistent.

Google’s current guidance specifically recommends keeping Merchant Center and Business Profile information up to date where relevant to generative AI search.

GEO and Brand Mentions

There’s another common misconception:

“Get as many brand mentions as possible.”

No.

The goal should be credible mentions, not artificial mentions.

Google explicitly warns against pursuing inauthentic mentions for AI visibility. Its current guidance says AI features can use information from blogs, videos, and forum discussions, but artificially seeking mentions isn’t a useful long-term strategy.

Better strategies include:

• Digital PR
• Expert interviews
• Original research
• Industry reports
• Podcast appearances
• Expert contributions
• Genuine reviews
• Community participation

The question isn’t:

“How many websites mention us?”

It’s:

“Which credible sources are talking about us, and why?”

How to Measure GEO

This is where the field is becoming more mature.

Traditional SEO reporting includes:

• Rankings
• Impressions
• Clicks
• CTR
• Organic traffic
• Conversions

GEO adds another layer.

You can track:

• AI citations
• AI mentions
• AI impressions
• Cited pages
• AI referrals
• Branded searches
• Conversion activity from AI traffic

Microsoft’s Bing Webmaster Tools now provides an AI Performance report with information such as citations, cited pages, and grounding queries.

Google also introduced dedicated Search Console reporting for generative AI visibility in June 2026, including data related to AI Overviews and AI Mode, with the reporting initially rolling out to a subset of sites.

This changes the conversation.

We’re moving from:

“I think ChatGPT mentions us.”

to:

“Which pages and topics are actually appearing in AI search experiences?”

Build a GEO Dashboard

I’d keep the first dashboard simple.

KPI	What to Monitor
AI citations	Number and trend
Cited URLs	Which pages are referenced
AI queries	Topics associated with visibility
AI impressions	Where available
AI referral traffic	Sessions from AI platforms
Organic traffic	Overall search performance
Branded searches	Brand demand
Leads	Qualified conversions
Revenue	Business impact

The most important part is the final connection:

Visibility → Traffic → Engagement → Conversion

Don’t celebrate 1,000 AI impressions if they generate nothing useful.

And don’t dismiss 50 highly qualified visits simply because the number is small.

A Better GEO Audit Framework

Instead of a generic SEO checklist, I’d audit five areas.

1. Retrieval Readiness

• Can search engines crawl the page?
• Is it indexed?
• Is the important content publicly accessible?
• Are important sections available as text?
• Are internal links strong?

2. Answer Quality

• Does the page answer the primary question?
• Is the answer easy to find?
• Are related questions addressed?
• Are explanations accurate?

3. Evidence

• Are important claims supported?
• Are statistics sourced?
• Are expert quotes legitimate?
• Are primary sources used?
• Is original research included?

4. Entity & Context

• Is the author clear?
• Is the organization clear?
• Are relationships between topics obvious?
• Does the internal-linking structure make sense?
• Is structured data accurate?

5. Multimodal Visibility

• Does the page have useful images?
• Are images properly optimized?
• Is video relevant?
• Could a diagram improve understanding?
• Are product visuals available where relevant?

This is a more useful GEO audit than simply checking whether a page has an XML sitemap.

The 30-Point GEO Checklist

Retrieval & Technical

• Page is indexable
• Page is crawlable
• Important content is available in text
• Robots.txt doesn’t block critical content
• Canonical is correct
• Internal links point to the page
• Page returns a valid response
• Important JavaScript content is accessible

Content

• Search intent is clear
• Main question is answered early
• Headings reflect actual questions/topics
• Sections are logically organized
• Content provides original value
• Claims are fact-checked
• Statistics have sources
• Expert quotes are authentic
• Outdated information is removed

Entities & Authority

• Author is identifiable
• Author expertise is demonstrated
• Organization is clearly described
• Products/services are clearly defined
• Internal relationships are strong
• Relevant external sources are cited
• Structured data matches visible content

Multimodal

• Relevant images are included
• Images have useful alt text
• Image filenames are descriptive
• Diagrams are used where helpful
• Relevant video is included
• Video content supports the written article

Measurement

• Search Console is configured
• Bing Webmaster Tools is configured
• AI visibility is monitored where available
• AI referral traffic is tracked
• Important landing pages are monitored
• Leads/conversions are attributed where possible

Common GEO Mistakes

Mistake 1: Treating GEO as a Hack

There’s no reliable:

“Add X and AI will cite you.”

The research doesn’t support that kind of universal formula.

Mistake 2: Writing for AI Instead of People

Google’s current guidance repeatedly emphasizes people-first, useful content.

That’s still the safest foundation.

Mistake 3: Creating Thousands of Fan-Out Pages

Don’t create a separate article for every possible variation of a query.

Google explicitly warns against generating pages primarily to manipulate rankings or AI responses.

Build meaningful topical coverage instead.

Mistake 4: Obsessing Over Tiny Content Chunks

There is no Google requirement to write 40-word paragraphs or split every article into tiny chunks for AI.

Clear structure is good.

Artificial chunking isn’t necessary.

Mistake 5: Treating Schema as a Citation Button

Schema can help describe content.

It doesn’t guarantee AI citations.

Mistake 6: Ignoring Images and Video

If your strategy is text-only, you’re leaving part of the generative search ecosystem untouched.

Google specifically highlights images and video as opportunities in AI search.

Mistake 7: Using Unsourced Statistics

If you use:

“Studies show…”

then link to the study.

If you don’t have the study, remove the sentence.

Mistake 8: Using Fake Expert Quotes

Never create a quote and attribute it to an expert.

If you paraphrase, say it’s a paraphrase.

If you quote, quote accurately and cite the source.

Mistake 9: Chasing Artificial Brand Mentions

Quality beats quantity.

Google explicitly cautions against inauthentic mentions.

GEO vs llms.txt

llms.txt has become one of the most discussed topics in AI SEO.

But it should be kept in perspective.

GEO	llms.txt
Broad optimization approach	Proposed machine-readable website file
Includes content	Can describe important resources
Includes technical SEO	Doesn’t replace technical SEO
Includes entities	Can provide additional context
Includes authority	Doesn’t create authority
Includes media	Doesn’t replace image/video SEO
Measurement possible	No guaranteed visibility benefit
Broader strategy	Optional experiment

Google’s current documentation says that llms.txt does not affect Google Search visibility because Google Search ignores it. Google says maintaining one won’t harm or help rankings in Google Search.

So if you’re using llms.txt, fine.

Just don’t build your entire GEO strategy around it.

What Actually Makes a Page Citation-Worthy?

If I had to reduce this entire guide to one section, it would be this.

A citation-worthy page usually has a reason to exist.

It doesn’t just rewrite what’s already available.

It might provide:

Original research

Something you discovered yourself.

First-hand experience

Something you’ve actually tested.

Unique data

Information competitors don’t have.

Expert interpretation

An informed explanation of complicated information.

Clear answers

Useful explanations without unnecessary fluff.

Strong evidence

Claims supported by credible sources.

Useful visuals

Diagrams, screenshots, videos or other media.

Clear context

Readers understand who created the information and why.

That’s the type of content I would want my own website to produce.

A Real-World GEO Example

Imagine you’re a SaaS company selling an analytics platform.

Your old blog article is:

What Is Conversion Rate Optimisation?

It contains 1,500 words.

It’s technically correct.

But 50 competitors have almost the same article.

Now imagine rebuilding it.

The new page includes:

A direct definition

A two-paragraph explanation.

Original data

Conversion patterns from your own customer base.

Screenshots

Examples from real dashboards.

A framework

Your own CRO methodology.

Expert commentary

Quotes from your internal CRO specialists.

Case study

A before-and-after example.

Video

A walkthrough of your CRO process.

Supporting articles

Links to:

• Landing page optimization
• A/B testing
• Conversion leaks
• Heatmaps
• Analytics

Sources

Links to relevant research and official documentation.

That’s not just “GEO-optimised”.

It’s simply a much better resource.

And that’s the point.

How to Build a GEO Content Cluster

For a website serious about GEO, I recommend using a pillar-and-cluster architecture.

Generative Engine Optimisation: The Complete Guide

Cluster Articles

• GEO vs SEO
• AI Citations
• AI Overviews
• ChatGPT Search
• Perplexity
• Gemini
• GEO Content Strategy
• Schema for AEO
• Entity SEO
• GEO Measurement
• GEO Tools
• llms.txt
• Multimodal GEO

Every cluster post should link back to the pillar within the first 100 words.

For example:

If you’re new to generative engine optimization, start with our complete guide to AI search visibility before moving into individual GEO tactics.

Then each cluster article should link naturally to two or three sibling posts.

The pillar should also link to the cluster articles through contextual sentences, not a generic “Related Posts” list.

GEO Content Architecture

If GEO is an important business topic, create a dedicated GEO Hub in the main navigation.

Don’t hide everything inside /seo/.

A structure could look like:

GEO Hub
│
├── Complete GEO Guide
├── GEO vs SEO
├── AI Citations
├── Google AI Overviews
├── ChatGPT Search
├── Perplexity
├── Gemini
├── GEO Content Strategy
├── Schema for AEO
├── Entity SEO
├── Multimodal GEO
├── GEO Measurement
└── GEO Tools

Use descriptive anchor text.

Never:

Click here.

Instead:

GEO vs SEO

or:

how to measure AI search visibility

This gives users and search engines useful context.

How Often Should You Update Your GEO Pillar?

Don’t update the date just to make the article look fresh.

Update the article when something meaningful changes.

For example:

• Google changes how AI Overviews work
• Search Console introduces a new AI report
• Bing adds a new AI measurement feature
• New research changes our understanding
• A major AI search platform changes
• A recommended practice becomes obsolete

Then update the content.

And update the last-modified date when the page has actually been meaningfully revised.

For a pillar page, freshness should reflect real editorial maintenance.

---

Frequently Asked Questions

What is generative engine optimization?

Generative engine optimization is the practice of improving content, technical accessibility, entities, authority, and supporting media so information can be discovered and potentially surfaced in AI-powered search experiences.

Is GEO the same as SEO?

No. GEO builds on SEO rather than replacing it. Google says its AI search experiences continue to rely on core Search ranking and quality systems.

Who coined the term GEO?

The 2023 research paper “GEO: Generative Engine Optimization” introduced the term as a framework for optimizing content visibility in generative engines.

Can GEO increase AI visibility?

Research suggests that content optimization can increase visibility in generative-engine responses. The original GEO study reported improvements of up to 40% in its experimental evaluation, although results varied by optimization strategy and domain.

Does adding statistics improve GEO?

Statistics can make content more informative and were among the optimization approaches examined in the original GEO research. However, you should not interpret the study as proof that adding statistics alone produces a fixed percentage increase in AI visibility.

Does schema markup improve GEO?

Schema can help search engines understand structured information and can support eligibility for certain search features. However, Google says there is no special Schema.org markup required for AI Overviews or AI Mode.

Do I need to break articles into tiny chunks for AI?

No. Google explicitly says there is no requirement to break content into tiny chunks for generative AI search. Focus on logical organization and reader-friendly sections instead.

Does llms.txt improve AI search visibility?

There is no Google Search benefit. Google currently says it ignores llms.txt and that maintaining the file neither helps nor harms Google Search visibility.

Are images important for GEO?

They can be. Google says generative AI search experiences can surface relevant images and videos and recommends supporting textual content with high-quality visual content where appropriate.

Does video matter for AI search?

Video can provide another way for users to discover and understand information. For tutorials, demonstrations, product explanations, and visual subjects, video can complement the written content and expand your search visibility opportunities.

How do I measure GEO?

Use available AI-search reporting alongside traditional SEO metrics. Google now provides generative-AI reporting in Search Console for eligible sites, while Bing Webmaster Tools provides AI Performance data including citations and cited pages.

Can I guarantee an AI citation?

No.

There is no legitimate GEO tactic that guarantees a citation from Google, ChatGPT, Perplexity, Gemini, or another AI search system.

The goal is to improve the quality, relevance, accessibility, and authority of your information—not manipulate an unknown ranking formula.

---

Final Thoughts: Stop Optimizing for the Citation

The biggest mistake I see in GEO is that people start with the wrong goal.

They ask:

“How do I get ChatGPT to mention my brand?”

I’d ask a different question:

“Why would an AI system want to use my content as a source?”

That question leads to much better work.

You start looking at evidence.

You improve the information architecture.

You publish original research.

You add first-hand experience.

You make important information easy to access.

You connect entities and related topics.

You use structured data accurately.

You add useful images and videos.

You measure actual AI visibility.

And you stop chasing every new GEO hack that appears on LinkedIn.

The original GEO research gives us evidence that optimization can influence visibility in generative search. But Google’s current guidance also makes something else very clear: there is no special collection of hacks that replaces good SEO.

The opportunity isn’t to make your website more “AI-like.”

It’s to make your website more useful, more authoritative, more understandable, and more valuable as a source.

That’s what I believe the next phase of SEO will look like.

Not SEO versus GEO.

Not humans versus AI.

But better information architecture, better content, better evidence, and better digital experiences that work across both traditional and generative search.

And if you’re building a GEO strategy today, I’d start with one thing:

Find the questions your audience actually needs answered—and become the best source for those answers.

Sources & Authority References

• GEO: Generative Engine Optimization — original 2023 research paper — foundational research introducing GEO and reporting experimental visibility improvements of up to 40%.
• Google Search Central: Optimizing your website for generative AI features — Google’s current guidance on AI Overviews, AI Mode, RAG, query fan-out, content, technical SEO, structured data, llms.txt, chunking and measurement.
• Google Search Central: AI features and your website — technical and content guidance for AI Overviews and AI Mode.
• Google Search Central: Top ways to succeed in AI Search — guidance on unique content, page experience, structured data, images, video and multimodal search.
• Google Search Central: Image SEO Best Practices — official image discoverability and optimization guidance.
• Google Search Central: Search Generative AI Performance Reports — current information about measuring generative-AI visibility through Search Console.