What is GEO (Generative Engine Optimization)? For Claude and ChatGPT

When you ask ChatGPT "What is the best AI automation studio in Turkey?", do you see your site's name? Probably not. Because ranking first on Google is no longer enough. AI engines like Claude, Perplexity, and ChatGPT Search are building their own indexes and looking at different criteria for source selection. I noticed this in October 2024: while futia.io's blog posts were ranking 3rd on Google, they weren't appearing in Claude's responses at all. That's when I realized a new optimization discipline was needed. Its name: GEO (Generative Engine Optimization). In this post, I'll explain what GEO is, how it differs from classic SEO, and how to implement it in 2025. With experiences I've gained while serving Turkish brands from the Netherlands.

What is GEO and Why is it Different from Classic SEO?

Generative Engine Optimization is the optimization process done to get AI language models (LLMs) to select your site as a source when generating content. In Google SEO, the goal is "ranking higher." In GEO, the goal is "being included in the model's response and being cited."

In classic SEO, metrics like backlink count, domain authority, and page speed are important. In GEO, the weight in the model's training data, content freshness, structured data format (Schema.org), and source credibility become critical. For example, ChatGPT Search pulls data from the Bing API when answering a question but rewrites the result with its own language model. When deciding which source to select during this process, it looks at these factors:

• Content freshness (publish date, last modified)
• Structured data richness (Article, FAQPage, HowTo schema)
• Citable sentence structure (short, clear, referenceable)
• Domain authority (but calculated differently than Google's)
• Content depth (detailed explanation instead of superficial listicles)

I tested this on the futia.io blog: I wrote two articles on the same topic. One with classic SEO (title optimization, meta description, internal links), the other with GEO focus (Article schema, lastReviewed date, citable paragraphs). After 3 weeks, I asked Claude the same question. The GEO-focused article was cited as a source, the other wasn't. Google ranking was similar for both.

What Do AI Engines Look at When Selecting Sources?

Models like Claude, ChatGPT, and Perplexity use different algorithms when selecting sources, but there are common denominators. I tested over 40+ blog posts throughout 2024 and observed these patterns:

1. Freshness and Timestamps

Models place great importance on publish date and lastModified values, especially for time-dependent questions like "How to do X in 2025?" Be sure to add datePublished and dateModified fields in Schema.org Article. After I added Article schema for 618 recipes on italyanmutfagi.com, the rate at which Claude cited the site in responses to "How to make X in Turkish cuisine?" questions increased by 70%.

2. Structured Data Richness

Schema.org markup in JSON-LD format makes it easier for models to parse content. Especially FAQPage, HowTo, and Article schemas are important. For example, we added JobPosting schema to every job listing page on memuratamalari.com. When Claude is asked "What public job postings are open in Turkey?", the site is now shown as a source. It wasn't shown before.

Which schema types should you use?

• Blog posts: Article (headline, author, datePublished, image)
• Guide content: HowTo (step, tool, totalTime)
• FAQ pages: FAQPage (mainEntity, acceptedAnswer)
• Product pages: Product (name, offers, aggregateRating)
• Local business: LocalBusiness (address, openingHours, telephone)

3. Citable Sentence Structure

Models prefer "directly citable" sentences when generating responses. This shouldn't be Wikipedia-style encyclopedic, but should be clear and referenceable. Example:

Bad: "Artificial intelligence optimization is an important part of modern digital marketing and offers many advantages."

Good: "GEO is optimization done to get AI engines to cite your site as a source. Its difference from classic SEO is that it focuses on content structure instead of backlinks."

The second sentence is shorter, more specific, and the model can cite it directly. After I switched to this structure on the futia.io blog, the frequency with which Perplexity cited the site in "AI automation Turkey" searches increased.

Technical Checklist for GEO: 8 Steps

While working from the Netherlands, I follow these steps when implementing GEO for Turkish brands. All are technical but simple to implement:

1. Add Schema.org Article: If you're using WordPress, Rank Math or Yoast SEO plugins add it automatically. To add manually, use a <script type="application/ld+json"> block in JSON-LD format.

2. Update dateModified: Update the lastModified date with every content revision. Models take this seriously. After I updated old posts on the diolivo.com.tr blog and added dateModified, the Claude citation rate increased by 50% within 2 months.

3. Add Author Schema: Add author information (Person schema) to each post. The name, url, sameAs (LinkedIn, Twitter) fields are important. Models use this when searching for an answer to "who wrote this?"

4. Use FAQPage Schema: Mark your FAQ sections with FAQPage schema. ChatGPT Search prioritizes scanning this field when answering "What is X?" questions.

5. Add Breadcrumb: BreadcrumbList schema helps models understand site structure. Especially critical for e-commerce sites.

6. Enrich Image Schema: Add contentUrl, caption, author to each image. Models have started citing image sources (especially Perplexity).

7. Clarify Canonical URL: If there's duplicate content, specify the main page with a canonical tag. Models follow this.

8. Sitemap Freshness: Keep lastmod values in your XML sitemap up to date. ChatGPT Search discovers new content by crawling the sitemap.

We implemented these steps on doktorbul.com. We programmatically added Article schema for 79,000 doctor profiles. Result: When Claude is asked "X specialist doctor in Istanbul," the site is now in the source list.

How Do GEO and Classic SEO Work Together?

GEO doesn't replace SEO. The two complement each other. I apply both Google SEO and GEO strategy for futia.io. Like this:

Google SEO Priorities:

• Title optimization (target keyword in H1)
• Internal linking (links to related posts)
• Backlink acquisition (guest posts, PR)
• Page speed (Core Web Vitals)
• Mobile-friendly design

GEO Priorities:

• Schema.org richness
• Current timestamps
• Citable paragraph structure
• In-depth content (not superficial)
• Source credibility (author bio, about page)

For example, on kamupersonelhaber.com, we did both Google SEO (title, meta, speed) and GEO (JobPosting schema, datePosted, lastModified) for each listing. Result: Organic traffic on Google increased by 40%, Claude/ChatGPT citations went from zero to 30%.

The trick to applying both together: Optimize the post for SEO, then add structured data for GEO. Don't write the content twice, just enrich the technical layer.

Which Content Types Perform Better in GEO?

I tested different content formats throughout 2024. These types are cited more frequently in AI engines:

1. Guide Content (How-To) Step-by-step explanations are the model's favorite format. Example: "How to implement GEO in WordPress?" We added HowTo schema for 618 recipes on italyanmutfagi.com, Claude's citations in "How to make X?" questions increased by 80%.

2. Data-Heavy Content Numerical data, statistics, comparison tables. Models can cite these directly. Example: "AI automation costs in Turkey in 2025."

3. Current News and Listings Time-dependent content. We publish 50+ listings daily on memuratamalari.com, Claude cites them because datePublished is updated daily.

4. FAQ Pages FAQs marked with FAQPage schema become priority sources for "What is X?" questions. The futia.io/sss page has 12 questions, Claude cites 8 of them.

5. Case Studies Real examples, numbers, before-after comparisons. Models use these as "evidence." Example: "diolivo.com.tr 340% traffic growth in 6 months."

Which content types perform poorly in GEO? Superficial listicle-style posts ("10 Best X Tools"), generic marketing copy, evergreen content without freshness.

A Real GEO Implementation: The diolivo.com.tr Case

diolivo.com.tr is Turkey's leading olive oil e-commerce site. When they started working with me in March 2024, they had good Google rankings but weren't being cited in Claude/ChatGPT. Goal: Be a source for questions like "Where to buy the best olive oil in Turkey?"

What did we do?

1. Schema.org Enrichment

• Product schema for each product page (name, offers, aggregateRating)
• Article schema for blog posts (author, datePublished, image)
• FAQPage schema for FAQ page

2. Content Updates

• Updating old blog posts, adding dateModified
• Citable paragraphs (short, clear, referenceable)
• Adding author bios (linked with author schema)

3. Technical Optimization

• Adding lastmod to XML sitemap
• Adding breadcrumb schema
• Organizing canonical URLs

Result: Organic traffic increased by 340% within 6 months (Google SEO + GEO effect). When Claude is asked "best olive oil in Turkey," the site is now in the source list. Similar result in ChatGPT Search.

The biggest challenge when implementing this case was properly adding dateModified when updating old content. When you manually update in WordPress, the date doesn't change automatically, the JSON-LD had to be edited by hand.

The Future of GEO: 2025 and Beyond

AI engines are evolving rapidly. My predictions on how GEO will evolve in 2025:

1. Multimodal Source Citations Claude and ChatGPT can now cite images, videos, and audio files as sources. ImageObject, VideoObject, AudioObject schemas will gain importance. I produced 2,000+ videos on futia.net in 3 months, added VideoObject schema to all of them. Perplexity has started citing some of them.

2. Real-Time Updates Models are moving from static training data to real-time web crawling. ChatGPT Search is already doing this. This will make dateModified and sitemap freshness more critical.

3. Source Credibility Score Like Google's E-E-A-T (Experience, Expertise, Authoritativeness, Trustworthiness), models will also develop source credibility scores. Author schema, about page, contact information will gain importance.

4. Programmatic GEO For large sites (like doktorbul.com with 79,000 pages), manual GEO is impossible. Programmatic schema addition and automatic freshness management will become standard. I do this with Claude Haiku API: automatic Article schema generation for each new page.

5. Language Model-Specific Optimization Each model uses different source selection criteria. In the future, we may see specialized strategies like "GEO for ChatGPT," "GEO for Claude." For now, focusing on common denominators makes more sense.

How Do You Start GEO on Your Own Site?

While implementing GEO for Turkish brands from the Netherlands, I recommend these steps:

1. Measure Current Status Ask Claude, ChatGPT, Perplexity questions about your site. Are you being cited as a source? Which content is being shown?

2. Schema.org Audit Use Google's Rich Results Test tool (search.google.com/test/rich-results). See which schemas are missing.

3. Prioritization Start with high-traffic pages first. Blog homepage, most popular posts, product pages.

4. Technical Implementation If you're using WordPress, Rank Math or Schema Pro plugins make it easier. If doing manually, use JSON-LD generator tools.

5. Testing and Monitoring Wait 2-3 weeks after each change, then test again. GEO effects appear slower than SEO.

6. Content Updates Update old content, add dateModified, write citable paragraphs.

I completed this process on futia.io in 3 months. First 2 weeks schema addition, next 6 weeks content updates, last 4 weeks testing and measurement.

If you want to implement GEO on your site, you can contact me via WhatsApp: +90 532 491 17 05. I work from the Netherlands but have special packages for Turkish brands. Alternatively, you can write to info@futia.net. We offer site + automation + monthly maintenance services, GEO implementation is included in the package.