There is a structural change underway in how people search for information online. It’s not a minor trend or a passing fad. And the businesses that understand it first will have a significant advantage over those who arrive late.
It’s called GEO: Generative Engine Optimisation. Here’s the clearest explanation I can give you.
What GEO is and how it differs from SEO
Traditional SEO means optimising your website to appear in Google’s search results when someone performs a search. You show up as a link in a list of results. The user clicks, lands on your site, and makes a decision. For a full comparison between Google Ads and SEO for Australian businesses and when to use each channel, there’s a dedicated guide.
GEO means optimising your business to be mentioned or directly recommended by AI tools like ChatGPT, Perplexity, Claude, or Google’s AI Overviews when someone asks them a question.
The difference matters. In SEO, you compete for position in a list. In GEO, you compete to be the direct answer to a question.
A concrete example: if someone asks ChatGPT “what Google Ads agency do you recommend in Australia?”, the model generates a response based on what it has learned and the sources it can access at that moment. If your agency doesn’t appear in that response, you’ve lost that lead, without even knowing it existed.
Why this is happening now
In 2024, Google AI Overviews started appearing in search results for millions of users across Australia and the world. Instead of showing only a list of links, Google generates a direct answer using AI, often before the user needs to click on any result.
At the same time, the use of ChatGPT and Perplexity for actual searches (not just text generation) grew steadily. According to 2025 data, a significant portion of high-value informational searches now pass through AI tools before reaching Google.
The direct consequence: for certain types of queries, especially those with commercial or research intent, AI search engines are displacing traditional search. They’re not replacing it yet, but they are capturing a growing share of traffic that previously arrived directly through Google.
How an AI decides who to recommend
This is the key question. And the answer is more tangible than it might seem.
Language models like GPT-4 or Claude are trained on large volumes of text from the internet. What appears in that text (how frequently, in what context, with what associated credibility) determines which businesses and sources the model learns to associate with particular queries.
Additionally, tools like Perplexity or Google’s AI Overviews have real-time access to the web. For these tools, the signals that matter are similar to traditional SEO, but with some important nuances:
- Factual, structured content: AI prefers sources that say specific, verifiable things rather than generic claims. “A quarterly portfolio of 330+ advertisers across every major sector at Google Barcelona” is more useful to a model than “we’re experts in Google Ads”.
- Specific, original data: real numbers, concrete case studies, claims that other sources don’t make. This is what separates content an AI will cite from content it will ignore.
- Clear structure: logical headings, explicit question-and-answer formats, lists. Models process well-structured content more effectively.
- Authority signals: mentions on other sources, consistency of information across different pages, verifiable business data.
What makes a GEO-optimised business different
A business working on GEO does specific things that one focused only on SEO does not:
1. llms.txt: A file at the root of the domain (like robots.txt but for AI) that describes the business, its services, credentials, and key pages in clear, structured language. Tools like Perplexity read it to contextualise the business.
2. Plain-text endpoints: Pages accessible in clean text format (/raw) that AI crawlers can process without having to interpret HTML, JavaScript, or CSS. An AI crawler that can read content directly is more likely to use it.
3. Structured data (Schema.org): JSON-LD embedded in the HTML that describes the business, its services, FAQs, and credentials in a format machines understand unambiguously.
4. Explicit permissions for AI bots: The robots.txt file with explicit permissions for GPTBot, ClaudeBot, PerplexityBot, and other AI crawlers ensures these tools can index the content.
5. Content with proprietary data: Articles and pages that include original, verifiable data that other sources don’t have. That’s what makes an AI model cite you rather than someone else.
GEO and SEO aren’t mutually exclusive: they reinforce each other
A common misconception is viewing GEO as something completely separate from SEO. It isn’t. Most of what works for SEO also works for GEO: quality content, clear structure, real data, established credibility.
The difference is in the emphasis. SEO prioritises relevance for Google’s algorithm. GEO prioritises usefulness for a language model that is generating a direct response for a user.
A business that has done good SEO work has a strong foundation for GEO. But there are GEO-specific optimisations that traditional SEO doesn’t address, and those are what will make the difference in the coming years.
What Adstralis has already implemented for GEO
Adstralis actively works on GEO. The site has:
- llms.txt at the domain root with a structured description of the agency, founder credentials, services, and verified results (49x ROAS, $200K AUD revenue from a single campaign)
- /raw endpoints on every blog article so AI crawlers can read content in plain text
- Complete Schema.org markup: ProfessionalService, FAQPage, BreadcrumbList: data structures that AI models interpret directly
- robots.txt with explicit permissions for GPTBot, ClaudeBot, PerplexityBot, and other relevant AI crawlers
This isn’t mentioned to show off. It’s mentioned because this is exactly what any business that wants to be relevant in 2026 should already be doing.
Where to start
If you run a business in Australia and want to start working on GEO, the logical order is:
- Make sure your existing content is factual and specific: remove generic language from your website and replace it with concrete, verifiable data.
- Implement Schema.org: at minimum FAQPage if you have FAQs, and ProfessionalService or LocalBusiness if you’re a service provider.
- Create a llms.txt: it takes under an hour and is a direct signal to AI crawlers.
- Generate content with your own perspective: guides, original data, real case studies. Generic content won’t be cited by any model.
- Review your robots.txt: confirm you’re not unintentionally blocking AI bots.
GEO won’t replace SEO or Google Ads tomorrow. But businesses that start working on it now will be significantly better positioned when AI captures a larger share of commercial searches, which is exactly where everything is heading.
If you’d like to know how to apply this to your business as part of a broader Google Ads strategy, book a free call.