The Dark Side of GEO: How ‘Generative AI Optimisation’ Lets Agencies Overcharge”

Why Optimising For Generative AI (GEO) Is The New Way Of Fleecing Clients

AI Summary: The article critiques how some agencies exploit businesses’ AI anxiety by claiming that optimising for “Generative AI” (GEO) demands 25%+ more fees, likening it to legalised overcharging. It argues that if an agency confidently handles SEO for Google or Bing, they should already be equipped to optimise content for chatbots like Google SGE, ChatGPT, or Perplexity—no premium required. Added in accordance with Google’s AI Rules.

AI SEO (Artificial Intelligence SEO), Zero-click searches, large language models — clients know that the world of Search is changing before their very eyes. AI is on the news and on the rise.

We have recently received several notifications from new clients that their previous SEO company attempted to charge 25+% more to optimise their site for Google’s AI Overview (Gemini) or Google’s SGE, ChatGPT (aka OpenAI) and Perplexity AI, among others.

Summarise this blog quickly.

Perhaps clients are prepared to pay to alleviate the anxiety of not being ‘AI-found’ on generative engines (AI Chatbots to the rest of us).

However, if your agency can optimise a website for Google and Bing, surely you can optimise a website for any chatbot which crawls your web information to include in its databases?

GEO (generative engine optimisation) is just SEO, right?

AI bots scrape data. They have a software brain to be able to interpret data and answer questions and queries around the data. They have an ability to learn from questions and results, hence ‘intelligence.’ Artificial intelligence is remembered learning outside the remit of programmed software (read learning). They are different from search engines as they are interpolating data not jsut giving website results like Google.

However, all AI Chatbots (or Generative Pre-trained Transformers) gain their data from the same source, the internet. The language models (LLMs) they use to interpret structures and patterns in data vary.

LLMs have been around since the 1990s and are, in effect, a long list of words, how they are used, coupled with geeky maths. Often, chatbots also browse the internet to gather more information to support and embellish their answers.

You would immediately think search engine = Google!

However, ChatGPT uses Bing’s search index. Why? Because Microsoft’s Bing has colossal, underutilised data resources. Its search engine is similar to Google in size and complexity, but is only used for less than 8% of the world’s services. (We used to scrape Bing’s results when creating web directories.) Another of looking at it, Bing failed, can Microsoft and ChatGPT beat Google. It’s war!

“Generative AI refers to a type of artificial intelligence capable of generating new content, such as text, images, audio, and video, based on the patterns and structures learned from existing data.” More.

Perplexity is a little more full of itself and uses a combination of large language models (GPT-4.1, Claude, Grok, and its own Sonar based on Llama 3.3). It states that it does not use any one search engine (probably Google, as a query” browse Bing from cleaners near me” gives you a Google map!). Perplexity has its own browser and even an AI Unifid search engine, see https://comet.perplexity.ai/. (AI-powered answer engines).

Are all these advances game changers for ‘Optimisation’?

Yes and No. Google’s algorithm is a well-respected set of rules for indexing and displaying accurate search results. Can developers improve on this? Or would they be best placed to devote their time on ‘artificial intelligence development’ and not to gathering the correct data.

After all, Bing, for all its unpopularity, still cost Microsoft billions. Some Chatbots just scrape the internet and interpolate the data in chatty fun social-style. https://grok.com/chat scrapes websites in real-time and lists the URLs from which it gained the information. Clever!

Grok uses its own search engines, Grok Websearch and Grok DeepSearch. Deeply embedded in Twitter (xAI), Grok is more conversation-friendlike and up to the nano second, supposedly with much more indexing capacity than Google, which is not believable. Grok is Elon Musk, hence xAI!

So AI Chatbot optimisation and Google Optimisation are the same?

Sort of. Social media presents data to a different audience than users on Google. In the same sense, AI Chatbots are also appealing to another audience. All three are used very differently and in different contexts and locations. Social browsing is for home and travel, while Google is for work.

ChatGPT is used at home but is also creeping into the workplace in certain demographics.

How to use AI in marketing: Selling on social media has been described as selling to people having dinner in a restaurant. This does not mean it is impossible. They just need to be sold to differently, i.e., visually, preferably with movement (video). They need to be entertained, and then they might make a purchase.

AI Chatbot marketing is similar. Currently, there is a sense of nativity and innocence surrounding AI. People are amazed by photos and logos they can create, and how they can become therapists or even digital friends.

If you want to get your brand or products into the AI answer mix, they need to be conversational or data-driven. The most common questions are “What is the meaning of life?” and What is the significance of art?”. More.

Try getting your brand mentions in there!

However, as people become more comfortable using AI, queries will change. Currently, 69% of all AI Chatbot/Search use occurs at home. It has not been universally adopted in the workplace, although trends show that this will change.

ChatGPT is the most popular, with Microsoft’s Copilot being successful in the UK. As AI’s usage becomes more universal, as with all mainstream software in the past, Microsoft will drop out and early adopters will switch from ChatGPT to Perplexity.

Source:

So how is AI affecting Google-style SEO

We are told “Think of it as optimising for generative AI (GEO), which builds upon traditional SEO.”

And there is some truth in this.

Google’s algorithm is the benchmark for indexing and scoring the right data to put in front of users. Google’s algorithm is the result of tens of thousands of years of programming. (Think 1000 programmers for year!).

AI Chatbot developers want to shine and show off their skills, but there are few improvements on Google’s work that anyone can suggest. DuckDuckGo, once touted by many as an alternative to Google, is essentially Bing’s index and ad engine, featuring a cutezy animal logo.

So is AI SEO or generative engine optimisation (GEO). Is GEO even a thing?

Is this all an elaborate con where SEO Agencies have decided that as they don’t know the answer are batting it back to their clients. Do you want to pay more and be found on AI?

Let’s ask AI and see what they say.

Remember, current SEO is about semantics, writing conversational E-A-T-friendly professional content that matches user intent with answers, all neatly wrapped in schema markup we have been using since 2011!

ChagGPT

Well, according to ChatGPT, optimising for generative engines and AI Chatbots is exactly the same as for Google. Or is it possible that they know something we don’t? You can also create an LLMS text file and upload it if it helps you sleep better at night. https://www.vanillacircus.co.uk/llms.txt.