🔥 The SEOs Diners Club - Issue #186 - Weekly SEO Tips & News

Hello fellow SEOs,

This week, we're tackling a truly critical topic that's been circulating in our industry. We need to put the widely discussed but fundamentally flawed concept of "AI Ranking" under the microscope. I'll walk you through how artificial intelligence actually works, how your brand can become visible in this new landscape, and most importantly, how you can achieve more effective results with what we call "context engineering."

If you're ready, let's embark on this exciting yet complex journey together.

🎯 The AI Ranking Myth and the Power of Context Engineering

Fellow SEOs, I'd like to address one of the most prevalent misconceptions circulating in digital marketing circles. The term "AI Ranking" is incredibly misleading, and we need to be crystal clear about this fundamental misunderstanding.

Large Language Models (LLMs) don't function like Google or traditional search engines. They're not information retrieval systems - they don't crawl websites, assign PageRank scores, or rank websites in any sense we understand. Grasping this fundamental difference is crucial for developing our future strategies effectively.

Think about it: How can you rank something that gives different answers to the same questions? You simply can't. LLMs are probabilistic, meaning variability isn't a bug - it's the fundamental nature of this technology. The entire concept of "rankings" achieved through consistency simply doesn't apply here.

Attempting to reverse-engineer AI results is impossible. Even research laboratories working on "interpretability" - trying to understand the reasoning behind machine learning system decisions or outputs - struggle with this challenge. Accepting this reality is the first step in developing sound strategies.

How AI Actually Works: Understanding the Core Mechanism

At the heart of generative AI lie Large Language Models (LLMs). LLMs have one singular task: predicting the next "token" for any given text. A token is a small piece of text from a word or word fragment, mathematically equivalent to approximately three-quarters of a word.

LLMs are trained on billions of documents during the pre-training process. These documents include web pages, social media sites, Reddit comments, and everything from Shakespeare to Stephen Hawking. During this process, the machine adjusts hundreds of billions of "knobs" and can subsequently generate the next word with mind-boggling accuracy.

But here's the critical point: LLMs are not deterministic. They don't always give the same answer. They deal with probabilities, not absolutes. This is an inherent characteristic of their nature, and understanding this feature determines how we work with them.

AI's Limitations and Ways to Overcome Them

AI is remarkable, but it has several important limitations:

1. Limited by Training Data: LLMs are constrained by their training data - they don't know everything. They have no access to information not present in their training datasets.

2. Hallucination: Sometimes AI "hallucinates," presenting false information as if it were true. This is particularly important when making critical decisions.

3. Frozen in Time: LLMs are frozen in time - they possess knowledge from when they were trained. Each conversation starts from scratch, somewhat like the movie "Memento" - they cannot form new memories.

To overcome these limitations, we provide AI with "context." This is where the concept of "context engineering" comes into play.

Context is King: The Most Important Skill of the New Era

In the AI world, context reigns supreme. The better context you provide to AI, the better results you'll achieve. This isn't just a technical detail - it's the fundamental determinant of your future success.

When you interact with an AI application (ChatGPT, Claude, Gemini, etc.), it places your query into what's called a "context window." Everything that enters the machine must fit within this context window.

HubSpot founder Dharmesh Shah's proposed PART formula helps us understand what goes into the context window:

P - Prompt: Everything you write must fit into the context window. This includes not just your question, but all instructions about how you want it to behave, what role it should take, and what kind of output you expect.

A - Archive: Chat history, long-term memories, everything needed to respond to the current prompt. This maintains the context of your previous conversations.

R - Resources: When we ask AI to analyze a PDF, look at a spreadsheet, or tell us what's in an image, these resources go into the context window.

T - Tools: We can provide LLMs with tools like browser access to the internet, database access, or API access to third-party systems. This allows LLMs to go beyond their historical limitations.

Modern LLMs can support context windows up to one million tokens, equivalent to approximately 750,000 words. This means you can provide massive amounts of information to AI simultaneously.

From Prompt Engineering to Context Engineering

The quality of results you get from AI depends on three fundamental things:

1. Model Quality: Use top-tier models like OpenAI's GPT-4, Anthropic's Claude, and Google's Gemini. Model quality determines the upper limit of results you can achieve.

2. Prompt Quality: Here's where the 60-30-10 Rule comes into play:

• Spend 60% of your time repeating working prompts

• 30% improving existing prompts (iteration)

• 10% experimenting with AI on things you've never tried

Meta Prompting technique is also valuable here: Give your own prompt to AI and say, "Can you make this prompt better? Ask me the questions you need, and I'll clarify." AI will significantly improve your prompt by asking about your target audience, length, sources, etc. This is a one-time investment that provides lifelong value.

3. Context Quality: We need to move from prompt engineering to the broader concept of "context engineering." This is about how to put the right things in the context window.

Custom Instructions: Tell AI who you are, what you do, how you want it to behave, what personality you want it to have, and what writing style you prefer. This is also a one-time investment that greatly improves your results.

MCP (Model Context Protocol): A powerful way to provide tools to LLMs in a standardized manner - think of it as the AI version of USB-C. This protocol connects AI applications with thousands of other applications, allowing you to bring all context directly to the LLM.

The Magic Prompt Trick: Clarifying Questions

After speaking with several prompt engineers, I discovered they all use a simple trick that significantly improves prompt results. Just one sentence, but the results are quite surprising.

One of the most common reasons that prompts give mediocre results is that you haven't provided the right context. But even if you provide AI with lots of context, you continue guessing what's most important, and so does AI. This can lead to "hallucination" or outputs different from what you actually want.

Instead, you can add this to the end of your prompts:

"Ask me clarifying questions to better understand my needs. Ask one question at a time and continue until you're confident you have enough information to provide a complete answer."

This works like magic. AI will guide you to clarify what you want, allowing you to get much more accurate and useful results.

How Does Your Brand Appear in AI? Proven Strategies

Since there's no such thing as AI ranking, how can brands appear in AI-powered searches or LLM outputs? The answer lies in what we know, but with a transformative approach: building authority and creating a solid foundation.

Instead of trying to reverse-engineer AI results (which you can't), focus on two proven approaches to increase your brand's visibility:

1. Building Brand Authority

Ensure your brand is consistently mentioned online, particularly strengthening your brand's positioning and message within the context where you want to be known. This is fundamentally about Public Relations (PR) and Content Marketing strategies.

Rather than chasing algorithmic mysteries you can't control, invest in what you can control: create extraordinary content, build genuine relationships, and develop real expertise.

The Power of Online PR: Media relations, collaborations with industry leaders, and influencer marketing ensure your brand name appears in the right places, in the right context. This increases your brand's recognition and credibility in the vast data pool that feeds LLMs.

Quality Content Marketing: Add value to your target audience through detailed blog posts, informative guides, research articles, and videos. Such content proves your industry expertise and naturally increases brand authority. High-quality, unique, and comprehensive content is perceived by LLMs as an important source of information.

2. Fundamental SEO Practices

SEO best practices still carry great importance. Because AI and LLMs use real-time search results to access current information. So, your website that ranks well on Google and has authority becomes a reliable source from which AI can draw information.

Technical SEO: Technical elements like your website's speed, mobile compatibility, crawlability, and indexability must be flawless. This ensures LLMs can easily access your content to use as a source.

Keyword Research and On-Page SEO: Identify keywords aligned with your target audience's search queries and use these keywords naturally in your content. Optimizing elements like title tags, meta descriptions, URL structures, and content hierarchy helps AI better understand your content.

E-E-A-T (Experience, Expertise, Authority, Trustworthiness): These principles that form the foundation of Google's quality guidelines also apply to AI. AI prefers to draw information from authoritative and trustworthy sources. Therefore, ensure your content reflects these four principles.

Conclusion: Human Intelligence + Augmented AI = The Future

AI isn't here to replace us. It's here to take over the repetitive, tedious parts of our work that don't bring us joy, encouraging us to focus on more meaningful things.

The paradox is beautiful: The better AI becomes, the more it enables us to be human. And being human isn't a flaw - it's our greatest feature.

As digital marketing professionals, we should use AI to test, clarify, and elevate our thinking, but never let it replace our thinking. Because no matter how smart AI becomes, humans win when it comes to emotional intelligence (EQ). AI is software - it has no emotions, no life experience, and feels nothing. You, however, have lived.

The future doesn't belong to AI, but to augmented intelligence with you. In this exciting period, rather than chasing misleading concepts like "AI Ranking" to increase your brand's online visibility, focus on solid strategies, building authority, and using AI intelligently.

🧠 Algorithm & Search Engine Updates

This week brought more activity in the search engine world, with Google updating its quality evaluator guidelines while once again demonstrating its sensitivity regarding YMYL content.

1. Google Updated Its Quality Rater Guidelines

Google made minor updates to its quality evaluator guidelines on September 11, providing clarifications on YMYL definitions and adding new examples for AI Overviews.

Action Plan: Emphasize the accuracy of your content, source reliability, and author expertise. Stick closely to E-E-A-T principles, especially for content falling under YMYL categories.

2. YMYL Now Includes Election and Civic Content

Google expanded the YMYL category with the "Government, Civics & Society" label. Election and voting information is now evaluated under YMYL.

Action Plan: If you're producing content about elections, voting procedures, candidate information, or local civic processes, be even more careful about accuracy and source citation.

3. Structured Data is Key to AI Visibility

Google, Microsoft, and OpenAI clearly state that structured data helps them better understand content. Now it's not just about content, but about the context of content.

Action Plan: Audit your website's structured data usage. Create a "content knowledge graph" to make your brand's digital footprint more understandable to AI.

4. Reddit Announced Pro Tools for Publishers

With 110 million daily active users, Reddit began offering specialized tools for publishers. Features like article insights, auto-import, and AI-powered community recommendations are in beta.

Action Plan: Monitor subreddits related to your brand. Reddit's spirit is authenticity and transparency - focus on adding value to the community rather than direct advertising.

🤖 From the AI World

1. 95% of ChatGPT Users Still Use Google

A striking fact: 95.3% of ChatGPT users visit Google, but only 14.3% of Google users use ChatGPT. This shows Google remains the default starting point.

Takeaway: LLMs aren't killing search - they're expanding and enhancing it. Don't neglect your fundamental SEO strategies.

2. ChatGPT Traffic is More Engaging but Lower Volume

According to Siege Media data, ChatGPT traffic is 1.73% more engaging than organic search. However, the volume is much smaller.

Takeaway: AI referrals bring high-quality visitors. It's critical that your content provides specific answers to specific questions.

3. Google AI Mode Added 5 New Languages

Google's AI-powered search experience expanded beyond English for the first time in 6 months, becoming available in Hindi, Indonesian, Japanese, Korean, and Brazilian Portuguese.

Takeaway: Google is determined to globally roll out AI-based search. Turkish support may be coming soon.

4. OpenAI Developing LinkedIn Rival Recruitment Platform

The OpenAI Jobs Platform will launch in mid-2026, targeting "perfect matches between company needs and what employees can offer."

Takeaway: AI competency certificates will become increasingly important. Consider shaping your career in this direction.

5. Anthropic's Claude Briefly Went Silent

A brief outage on September 10 reminded us how complex and sensitive AI systems are. Users' jokes about "having to use our brains" show how dependent we've become on these technologies.

💡 Practical SEO Tips

1. Move from Prompt Engineering to Context Engineering

As mentioned in our main article, the quality of results you get from AI depends on three things:

• Model Quality: Use top-tier models

• Prompt Quality: Apply the 60-30-10 rule (60% repetition, 30% improvement, 10% experimentation)

• Context Quality: Use the PART formula

Practical Tip: Do meta-prompting. Give your own prompt to AI and ask "Can you make this prompt better?"

2. Focus on Q&A Format

Post-ChatGPT, users make longer, more conversational queries.

Practical Tip: Filter for question words like "how," "what is," "when," and "best" in Google Search Console. Create content that provides clear, comprehensive answers to these questions.

3. Guide AI with Structured Data

Go beyond basic schemas. Define your brand's competency with AboutPage, ProfilePage schemas.

Practical Tip: Use the sameAs property to connect your website with your social media profiles. This creates a digital identity card that clarifies who you are to AI.

4. Sample Prompts

Content Idea Generation: "You are an experienced SEO strategist. Create 10 blog post ideas for [industry] targeting [target audience], revolving around '[keyword]'. Include description, title, and competition level for each idea."

Media List for Online PR: "You are a PR expert. List 7 journalists/influencers at [field]-focused publications for [product/service] launch. Include profile, sample work, and contact information."

Remember: Add the magic sentence to the end of every prompt: "Ask me clarifying questions to better understand my needs."

👋 Closing

We've reached the end of this week's SEOs Diners Club. As you can see, rather than chasing misleading concepts like "AI Ranking," we need to focus on solid fundamentals.

Remember:

• There's no such thing as an AI ranking

• Context is queen

• Brand authority and fundamental SEO remain the most important strategies

• Move from prompt engineering to context engineering

If you need a strategic partner to navigate your brand through these winds of change, we at Stradiji are here to guide you in SEO, content marketing, and digital PR.

The future doesn't belong to AI, but to people with augmented intelligence who use AI wisely.

Until next week with new developments,

Best regards,

Mert Erkal,

Founder, Stradiji - AI-Driven SEO Consultancy

Special Offer from Our Sponsor

Semust: Smarter SEO Recommendations for Better Organic Traffic

Managing your entire SEO workflow just got easier.

In this GEO era, where automation is king, you still need reliable tools to track, analyze, and optimize your SEO efforts. That's where Semust comes in.

Semust combines powerful SEO tools in a single platform:

• Google Rank Tracker - Monitor your keyword positions with precision

• Technical SEO Checker - Identify and fix technical issues fast

• SERP Analyzer - Understand what's ranking and why

• Complete workflow management - From audit to optimization

Exclusive Offer for SEOs Diners Club Readers:

Get a FREE one-month Business Plan (normally $35/month) to supercharge your SEO workflow.

How to claim: Email: [email protected] Subject: "SEOs Diners Club - Free Month"

Learn more about Semust | View pricing plans 

Want to Support This Newsletter?

I love sharing the latest developments and strategies from the SEO world with you. If you find my content useful, you can support me by buying a coffee. ☕ Click 'Buy Me a Coffee' to contribute to knowledge sharing. Let's achieve more together!

© 2022-2025 SEOs Diners Club by Mert Erkal