What Is Generative Engine Optimisation (GEO)?

Generative engine optimisation, often shortened to GEO, is the practice of improving how you appear in answers written by AI tools such as ChatGPT, Gemini, Perplexity, Claude and Google’s AI overviews.

SEO is about ranking links. GEO is about being understood, cited and summarised correctly when there is no list of ten blue links, only a paragraph that sounds like it knows you.

Why this sits inside reputation work
Three-quarters of adults now use AI tools to research people and brands. Those tools do not ring you if they get it wrong. They pull from Wikipedia, news, profiles, forums and whatever else they have been trained on or can retrieve, then they write it up with confidence.

For a person or a company, that summary is a first impression. If it is outdated, lopsided or invented, it can travel further than a single Google result, because the reader is not shown the messy context.

That is why GEO is not a separate gimmick from online reputation management. It is the same job, aimed at a new place people look.

GEO, SEO and AEO in plain terms
You will also see AEO, answer engine optimisation. People use the labels loosely. The useful distinction is:

SEO: helping the right pages rank on Google and Bing
AEO: helping your content be the answer to a specific question (including featured snippets and AI overviews)
GEO: helping generative tools describe you accurately, with sources they are willing to trust
In reputation work you need all three. Ranking a page does not automatically fix a ChatGPT biography. A clever AI tweak does not help if page one is still a problem article from 2016.

How AI decides what to say about you
Models differ, and they change. In practice they lean on:

Wikipedia, when a page exists
established news and reference sites
consistent facts repeated on strong, crawlable pages (your site, major profiles, filings)
whatever is easy to retrieve for your name this week
They do not know your side of the story unless it is published somewhere they can use. They do not care that you have moved on from a role, a case or a product line unless newer, clearer sources exist.

What GEO work actually looks like
A serious GEO piece of work is not stuffing “ChatGPT” into blog posts. It looks more like this:

Audit. Ask the main tools the questions a client, journalist or recruiter would ask. Save the answers. Note errors, omissions and tone.
Sources. Work out where those answers are coming from. Wikipedia, an old profile and a news piece will usually outweigh a new blog.
Fix the record. Strengthen accurate pages you control. Correct or contextualise what you can. Build the citations AI is more likely to reuse: clear bios, sourced explainers, proper Wikipedia process where you qualify.
Keep the facts consistent. Name, role, company, dates and spelling should match across your site and major profiles. Models punish messy entities.
Retest. Answers drift. Check again after you publish, and on a cycle, not only in a crisis.
What GEO cannot do
It cannot force ChatGPT to use your preferred wording. It cannot hide lawful journalism. It cannot replace Wikipedia’s rules with a shortcut.

Anyone selling “we will get you into ChatGPT this month” is selling a result nobody fully controls. Influence the sources. Measure the summaries. Stay honest about the rest.

A 20-minute check you can do today
Open ChatGPT, Gemini, Perplexity and, if you use it, Claude.
Ask: “Who is [your name]?” and “What is [your company]?”
Ask one commercial question a client might ask, such as “best online reputation management UK” if that is your world.
Write down what is right, what is missing, and what is simply false.
If that file would embarrass you in a first meeting, you have a GEO problem, even if Google still looks “fine”.

Next Chapter audits what Google and AI tools currently say, then works on the sources those systems actually use, including search, profiles and Wikipedia where it is appropriate. If you want that done confidentially, get in touch.