Something has shifted in how people find information online. A growing number of searches — especially research-stage, consideration-stage queries — are now answered directly by AI. ChatGPT. Perplexity. Google’s AI Overviews. Instead of a list of ten blue links, users get a synthesised answer, often with a handful of cited sources.
If your brand isn’t one of those sources, you’re invisible in an increasingly important part of the search landscape.
This is what Generative Engine Optimisation — GEO — is about.
What GEO actually means
GEO is the practice of making your content readable, credible and citable by AI-powered search tools. It’s related to traditional SEO but distinct from it in important ways.
Traditional SEO is about ranking in a list. GEO is about being cited in an answer.
To be cited, AI tools need to be able to do three things with your content:
- Find it — your site needs to be crawlable and indexable
- Read it — your content needs to be structured in a way AI can parse
- Trust it — your site needs signals that indicate authority and credibility
Most websites fail on at least one of these. Many fail on all three.
What we found when we audited a SaaS client
A global SaaS client came to us after noticing that competitors were appearing in AI-generated answers while they weren’t. An audit found an AI visibility health score of 65% — below industry benchmark — alongside over 2,600 technical and content issues.
The problems included duplicate tags, broken images, no structured content and no keyword mapping. The site couldn’t be confidently read, understood or cited by AI tools.
We rebuilt the technical foundation and restructured the content from the ground up. Within five months:
- AI health score up from 65% to 86%
- Total site errors down 78%
- Overall site health up from 75% to 92%
The practical steps
1. Fix the technical foundation
Before any content work, the site needs to be technically sound. That means:
- Eliminating duplicate titles and meta descriptions
- Fixing crawl errors and broken links
- Reducing crawled pages to protect crawl budget
- Rebuilding internal linking so AI can navigate the site logically
In the audit above, we cut crawled pages from 2,911 to 1,701. That’s not about hiding content — it’s about making sure every page that gets crawled is a page worth crawling.
2. Add an llms.txt file
This is a relatively new but important step. An llms.txt file is a plain text document that sits at the root of your domain and tells large language models what your site is, what it covers and how it should be understood.
Think of it as a robots.txt file, but for AI rather than search engine crawlers. It helps AI tools understand your brand, your expertise and your content before they start reading individual pages.
3. Restructure content around real questions
AI tools are trained to answer questions. If your content doesn’t directly answer the questions your customers are asking, it won’t get cited.
This means moving away from broad topic pages towards specific, question-focused content. It means writing in a way that’s clear, direct and easy to extract. And it means building FAQ content that’s explicitly designed for AI citation — questions phrased the way real users ask them, with concise, accurate answers.
4. Complete keyword mapping
Every page on your site should be mapped to a specific intent. When AI tools encounter a page, they need to understand immediately what question it answers and for whom.
If your pages are broadly topical rather than specifically intent-mapped, they’re less likely to be cited for any specific query.
5. Fix alt text and anchor text
AI tools use image alt text and link anchor text to understand context. If your images have generic alt text (“image1.jpg”) and your links say “click here,” the AI gets no useful signal about what those elements relate to.
Rebuilding alt text and anchor text for clarity is unglamorous work, but it has a meaningful impact on AI readability.
Is this instead of SEO or in addition to it?
In addition to. The fundamentals of traditional SEO — technical health, quality content, authority signals — are also the fundamentals of GEO. You’re not starting over; you’re extending.
The distinction is in how you think about content. SEO asks “how do I rank for this term?” GEO asks “how do I get cited as the answer to this question?” Both matter. The sites that will perform best over the next few years are those that optimise for both.
Where to start
If you’re not sure whether your site is AI-ready, the best first step is an audit. Look at your technical health, your content structure and your crawl data. The problems are usually visible quickly — and the fixes, while not instant, are straightforward.
If you’d like someone to do that audit for you, that’s exactly what a discovery call is for.