Why does it feel like the “rules” of AI search optimisation change every time we talk about it on this podcast? Probably because they do!

Since airing our last episode, we ran an AI citation study of 11,000+ citations across 1,000+ questions, and now we want to walk you through the findings to see which sources AI systems actually reference, and what that might mean for your content strategy.

We’ll cover the patterns we observed, such as the role of your off-site brand reputation, how different AI platforms and markets cite very different sources, and why small wording changes in prompts create huge shifts in the citation landscape.

We also bust a few myths:

• That AI mostly trusts “neutral editorial” sources
• That Reddit and user-generated content dominate citations
• That more backlinks automatically mean more AI citations
• That optimising for Bing is the shortcut to ChatGPT visibility

Treat this as a snapshot in time. It’s not a set of permanent truths, but one that gives you concrete hypotheses to test on your own site and topic clusters.

What You’ll Learn:

• The median cited passage length and why “self-contained” matters more than clever writing
• How heading structure and section purpose influence what AI pulls from your pages
• Why long and heavily formatted pages show up often, but aren’t necessarily the cause for more citations
• How review platforms make up 36% of brand-reputation citations (and what that implies for B2B)
• Why ChatGPT and Gemini share only ~5% of cited domains on average
• How localisation changes AI source lists far more than it changes classic SERPs
• Why tracking related prompts, not just one seed query, is essential to understanding your AI visibility
• Which commonly repeated GEO “truths” our data did not support

Resources Mentioned

• AI Citation Study
• Ep 5.2: What Happens to Keyword Research in the Wake of GEO?
• GeekyIQ

#AI #GEO #AISearch #ChatGPT #Gemini #Google #Bing

Get more SEO Unfiltered! Follow us on Facebook, Instagram, and X @GeekyTechGeeks for all things SEO/GEO, and advertising related—and while you're at it, why not subscribe, like, and follow us on Apple Podcasts, Spotify, Stitcher, or wherever you listen to your favourite shows.

We are now on YouTube! https://www.youtube.com/@geekytechgeeks

For more digital marketing insights, connect with us on LinkedIn: https://uk.linkedin.com/company/geeky-tech

Have any questions you want answered on the show? Email us at team@geekytech.co.uk.

Thanks for listening 🤓

GEO reporting is still evolving, but that doesn’t mean marketers can’t build a credible story around performance. In this episode of SEO Unfiltered, we explore how to benchmark AI visibility, track the right metrics, and pair emerging GEO signals with more concrete evidence from search and referral data.

In This Episode, We Look At:

• How to benchmark AI visibility using branded and unbranded prompts.
• Which GEO metrics are useful, and which ones need to be treated with caution?
• Why AI visibility should be paired with supporting evidence from GA4 and GSC.
• How to improve brand visibility across LLMs through content quality, brand consistency, and off-page authority.
• How to present concrete GEO progress.

Resources Mentioned

• Episode 5.3: AI Visibility Metrics: Are They All Bullshit?
• Ep 5.2: What Happens to Keyword Research in the Wake of GEO?
• GeekyIQ

Example Prompt for AI

Help me create a benchmark set of 30 AI visibility prompts for [brand name] in [industry]. I want the set to include branded, competitor-branded, and non-branded prompts so I can compare how often our brand appears in AI responses versus its competitors.

Please group the prompts by intent, such as discovery, comparison, evaluation, sentiment, and problem-solving.

Make the prompts realistic, varied, and reusable over time so they can be tracked consistently across AI tools.

Also explain briefly why each prompt matters and how it helps measure visibility.

Episode Takeaway

The goal of GEO reporting is not to prove that one AI mention caused one click. The goal is to show leadership a broader pattern: AI visibility is improving, search demand is moving, and your overall marketing strategy is building traction in the right places.

Want to Keep Listening?

If you found this episode useful, go back to the previous episode on AI visibility metrics for a deeper look at the measurement side of GEO. Then use this episode as the practical follow-up for turning those metrics into a reporting story your boss can actually understand.

See you in September!

Get more SEO Unfiltered! Follow us on Facebook, Instagram, and X @GeekyTechGeeks for all things SEO/GEO, and advertising related—and while you're at it, why not subscribe, like, and follow us on Apple Podcasts, Spotify, Stitcher, or wherever you listen to your favourite shows.

We are now on YouTube! https://www.youtube.com/@geekytechgeeks

For more digital marketing insights, connect with us on LinkedIn: https://uk.linkedin.com/company/geeky-tech

Have any questions you want answered on the show? Email us at team@geekytech.co.uk.

Thanks for listening 🤓

AI visibility tools are everywhere and they’re making big promises. But how much of what they’re selling is actually useful, and how much is just dressed-up guesswork? And is guesswork always bad?

In this episode, we break down the most common AI visibility metrics, what they really measure, and how marketers should (and shouldn’t) use them.

🎯 What You’ll Learn

• Why AI visibility scores are not based on first-party model data
• The difference between proxy metrics vs. real business outcomes
• How to interpret AI Share of Voice without turning it into a vanity metric
• Why citation count alone can be misleading
• The reality of AI brand sentiment (and why it’s often overhyped)
• How to actually use AI visibility data to inform strategy and content decisions

🔍 Key Topics Covered

AI Visibility Score: Insight or Illusion?
AI visibility scores are built on prompt testing, not actual model data. That makes them directional—not definitive. Useful for trends, risky as a KPI.

AI Share of Voice (SoV)
A helpful comparative metric, but still based on sampled prompts. Without context (like accuracy or sentiment), the number alone doesn’t tell the full story.

Citation Count vs. Citation Rate
Being mentioned isn’t the same as being cited. If AI isn’t linking to you, it may not see your content as a trusted source.

Brand Sentiment in AI
Most AI systems default to neutral or softened language, making sentiment less actionable than it sounds.

⚠️ The Big Takeaway

• AI visibility metrics aren’t useless, but watch out how they're being sold.
• They should guide your strategy, not dictate it.
• The real goal isn’t to “win” a visibility score. It’s to become a source AI systems consistently trust, cite, and recommend.

🧠 Practical Advice

• Focus on trends over snapshots
• Combine AI metrics with SEO data and real performance metrics
• Look beyond dashboards and prioritise actionable insights
• Don’t drop SEO fundamentals while chasing AI visibility

🛠️ And introducing.....

GeekyIQ: Geeky Tech's new AI Brand Intelligence Platform

• Track visibility, uncover topic gaps, and get prioritized recommendations to improve your presence in AI-generated responses.

Start your free trial: https://geekyiq.com

💬 Join the Conversation

We’ve officially entered our video podcast era 👀
Do you prefer watching or listening? Let us know.

Get more SEO Unfiltered! Follow us on Facebook, Instagram, and X @GeekyTechGeeks for all things SEO/GEO, and advertising related—and while you're at it, why not subscribe, like, and follow us on Apple Podcasts, Spotify, Stitcher, or wherever you listen to your favourite shows.

We are now on YouTube! https://www.youtube.com/@geekytechgeeks

For more digital marketing insights, connect with us on LinkedIn: https://uk.linkedin.com/company/geeky-tech

Have any questions you want answered on the show? Email us at team@geekytech.co.uk.

Thanks for listening 🤓

In this episode of SEO Unfiltered, we revisit one of SEO’s most fundamental tasks—keyword research—and explore how it’s evolving under the rise of GEO.

Host Genny cuts through the noise of “SEO is dead” hype to show how you can adapt traditional keyword strategies for AI-driven search engines like ChatGPT, Claude, and Gemini—without abandoning what already works.

If you’ve been wondering how to align your SEO workflows with emerging AI-driven search behaviour, this one’s for you.

Key Topics Covered

• What GEO (Generative Engine Optimisation) actually means and how it affects search visibility.
• Why keyword research isn’t dead—it’s just evolving.
• How users interact differently with generative tools vs. traditional search engines.
• The new mindset: focusing on search intent and problem-solving, not just exact-match phrases.
• A practical workflow for AI-inclusive keyword research using a real-world example.
• How to prioritise keywords and prompts based on value, opportunity, and visibility gaps.
• Building pillar and supporting pages that reinforce topical authority for both SEO and GEO.
• Why interlinking your clusters boosts crawlability, rankings, and AI citation potential.

Resources Mentioned in This Episode:

• Episode 5.1: “How to E-E-A-T Your Content for Better AI Visibility”

Get more SEO Unfiltered! Follow us on Facebook, Instagram, and X @GeekyTechGeeks for all things SEO/GEO, and advertising related—and while you're at it, why not subscribe, like, and follow us on Apple Podcasts, Spotify, Stitcher, or wherever you listen to your favourite shows.

We are now on YouTube! https://www.youtube.com/@geekytechgeeks

For more digital marketing insights, connect with us on LinkedIn: https://uk.linkedin.com/company/geeky-tech

Have any questions you want answered on the show? Email us at team@geekytech.co.uk.

Thanks for listening 🤓

Genny is back for season 5 with a solo rant/pep talk about E-E-A-T, AI, and why geologists everywhere probably want their acronym back. In this episode, we dust off Experience, Expertise, Authoritativeness, and Trustworthiness and give the whole framework a shiny new GEO/AI visibility twist.

What this episode is about:

• How E-E-A-T translates into actual, real-world content instead of fluffy theory.
• Why authority still matters, but the way AI systems understand and score that authority is changing.
• How Retrieval-Augmented Generation (RAG) works in plain English, and what that 0.6 “confidence threshold” means for whether your content is visible in AI answers… or invisible.

In 2026, E-E-A-T means...

• Experience: Content that proves you’ve actually done the thing—case studies, test results, lived stories—not just reworded someone else’s blog.
• Expertise: Credentials, track record, and depth of knowledge that show you’re the person people (and robots) should listen to, especially for YMYL topics.
• Authoritativeness: Other trustworthy people and brands cite you, link to you, and generally treat you like the grown-up in the room.
• Trustworthiness: All the “this site doesn’t feel shady” signals—technical, legal, reputational, and editorial—that give Google and AI the confidence to surface your content.

One word that keeps coming up in this episode is schema

• Review, HowTo, Person, Organisation, and Review schema all help translate your credibility into machine-readable signals.
• If markup makes your eyes glaze over, we strongly suggest looping in your SEO team or developer and treating schema as non-negotiable infrastructure.

Get more SEO Unfiltered! Follow us on Facebook, Instagram, and X @GeekyTechGeeks for all things SEO/GEO, and advertising related—and while you're at it, why not subscribe, like, and follow us on Apple Podcasts, Spotify, Stitcher, or wherever you listen to your favourite shows.

We are now on YouTube! https://www.youtube.com/@geekytechgeeks

For more digital marketing insights, connect with us on LinkedIn: https://uk.linkedin.com/company/geeky-tech

Have any questions you want answered on the show? Email us at team@geekytech.co.uk.

Thanks for listening 🤓

In this redux episode, we revisit our original 5 Bad Habits episode from March 2023—a time when ChatGPT had just been launched and digital marketing was forever changed. Since then, SEO has transformed with the rise of AI-driven search and generative engines and their effect on how people find and engage with content online.

In this episode, we cover topics like...

• Why resisting change is the biggest marketing mistake you could possibly make.
• How embracing AI-inclusive SEO practices like GEO/LLMO can future-proof your marketing.
• The importance of having a strong relationship with your data — tracking the right metrics, consolidating sources, and making data accessible using AI analytics.
• Why overlooking your marketing strategy, especially one aligned with customer intent and competitor insights, can cost you dearly.
• The risks of inconsistent branding and messaging that confuse your customers and weaken your professional image.
• How giving AI full control of your content without human oversight can harm your brand’s credibility and SEO performance.

As a little bonus, we also discuss why blocking LLMs from scraping your site is a counterproductive move, given the evolving AI landscape and the visibility it can bring.

Happy listening, marketing geeks! Here’s to mastering 2025 and beyond.

Get more SEO Unfiltered! Follow us on Facebook, Instagram, and X @GeekyTechGeeks for all things SEO/GEO, and advertising related—and while you're at it, why not subscribe, like, and follow us on Apple Podcasts, Spotify, Stitcher, or wherever you listen to your favourite shows.

We are now on YouTube! https://www.youtube.com/@geekytechgeeks

For more digital marketing insights, connect with us on LinkedIn: https://uk.linkedin.com/company/geeky-tech

Have any questions you want answered on the show? Email us at team@geekytech.co.uk.

Thanks for listening 🤓

In this episode, we dive deep into the most popular Large Language Models (LLMs) available for marketers. With AI tools becoming ubiquitous across every industry, this episode breaks down which LLMs are best suited for specific marketing needs.

Which LLMs you might ask?

1. ChatGPT 5

Best For:

• Brainstorming ideas
• Creating social media posts
• Generating structured data and schema markup
• Meta tag creation
• Email drafting
• Long-form content (with human editing)
• Image and code generation

Limitations:

• Can hallucinate (make up information)
• Struggles with basic math and accurate formulas
• Tends to repeat points in long-form content
• May use training data instead of web search for recent topics
• Quality depends heavily on prompt specificity

2. Gemini (Google)

Best For:

• Users embedded in Google Workspace ecosystem
• Image processing
• Real-time research and fact-checking
• PDF analysis
• Academic research

Considerations:

• Less creative than ChatGPT
• Always-present assistant feature
• Better for research, not ideal for creative content

3. DeepSeek

Best For:

• Developers and coders
• Mathematical accuracy
• Building spreadsheets and tables
• High-logic tasks
• Structured output

Limitations:

• Text-only (no images, audio, or video)
• Lacks artistic flair
• Not ideal for creative marketing content
• Better suited for technical users

4. Perplexity

Best For:

• In-depth research
• Fact-checking recent events
• Academic and professional research
• Getting cited, authoritative sources

Standout Features:

• Shows when sources were last updated
• Step-by-step breakdown of reasoning
• Allows you to choose search sources
• Suggests follow-up searches
• Less "sycophantic" than ChatGPT

5. Claude

Best For:

• Long-form content (whitepapers, research papers, e-books)
• Creative and expressive writing
• Tone and style switching
• Processing large amounts of text (high token capacity)

Limitations:

• Struggles with complex reasoning and layered topics
• Output can be generic without detailed prompts
• May filter contextually appropriate information if deemed potentially harmful
• Requires skilled prompting for best results

Get more SEO Unfiltered! Follow us on Facebook, Instagram, and X @GeekyTechGeeks for all things SEO/GEO, and advertising related—and while you're at it, why not subscribe, like, and follow us on Apple Podcasts, Spotify, Stitcher, or wherever you listen to your favourite shows.

We are now on YouTube! https://www.youtube.com/@geekytechgeeks

For more digital marketing insights, connect with us on LinkedIn: https://uk.linkedin.com/company/geeky-tech

Have any questions you want answered on the show? Email us at team@geekytech.co.uk.

Thanks for listening 🤓

We remember back in the day when SEO and PPC were just about the only abbreviations we had to worry about, but then came along AI, and boy oh boy, have things gotten just a little bit more alphabetty.

In this episode, Genny’s going to run through the list of the most common AI-related terms that marketers most likely encounter on their mission to maximise their visibility on AI-powered search.

In case you want spoilers, here’s the list:

AI (Artificial Intelligence)

• Any technology that simulates human intelligence in machines that allows them to perform tasks like recognise patterns, make predictions, and understand language.

GenAI (Generative AI)

• A type of AI that can generate new content, (i.e., text, images, music, or videos) based on its training data. Examples include ChatGPT, MidJourney, and Deepfakes.

GEO (Generative Engine Optimization)

• This is the process of optimising web content to maximise your brand visibility in generative AI systems (like ChatGPT) by ensuring your content gets selected as answers in AI-generated responses.

AEO (Answer Engine Optimization)

• Synonymous with GEO, AEO focuses on optimizing content so that it appears in AI-generated answers, particularly for Answer Engines.

AIO (AI Overviews)

• AIO refers to AI Overviews, those helpful content summaries at the top of Google’s search results, which include a short overview with relevant links.

LLM (Large Language Model)

• A type of AI that’s trained on vast amounts of text data that can process and generate human-like language. LLMs like GPT-3 can answer questions, generate text, and more, but they are based on fixed training data (they don’t learn in real time).

NLP (Natural Language Processing)

• This is a subset of AI that’s focused on helping machines understand, interpret, and generate human language, e.g., speech recognition, text analysis, and machine translation.

ML (Machine Learning)

• Another subset of AI where machines learn from data to recognize patterns and make predictions.

DL (Deep Learning)

• A part of machine learning that uses neural networks with multiple layers (hence the term "deep") to process and understand complex data.

AGI (Artificial General Intelligence)

• A type of AI that can learn and apply knowledge across various domains, just like humans. Unlike narrow AI, which excels at specific tasks (e.g., chatbots), AGI can perform a wide variety of tasks with human-like reasoning.

Hallucinations (AI Hallucinations)

• In this context, hallucinations refer to instances where the model generates incorrect or totally made up information that appears convincing.

GPT (Generative Pre-trained Transformer)

• A type of LLM (like ChatGPT), that uses transformer architecture to generate human-like text based on large-scale pretraining. GPT can generate, summarise, or answer questions based on data it has been trained on.

Remember, we’re at the very start of our AI journey, so don't let information overwhelm get in the way of AI adoption!

Happy Listening 🎧

Get more SEO Unfiltered! Follow us on Facebook, Instagram, and X @GeekyTechGeeks for all things SEO/GEO, and advertising related—and while you're at it, why not subscribe, like, and follow us on Apple Podcasts, Spotify, Stitcher, or wherever you listen to your favourite shows.

We are now on YouTube! https://www.youtube.com/@geekytechgeeks

For more digital marketing insights, connect with us on LinkedIn: https://uk.linkedin.com/company/geeky-tech

Have any questions you want answered on the show? Email us at team@geekytech.co.uk.

Thanks for listening 🤓