The definition
What is generative engine optimization (GEO)?
Generative engine optimization, or GEO, is the work of improving how a business appears in answers written by generative AI systems such as ChatGPT, Gemini, Claude, Perplexity and Google AI Overviews. The name comes from a research paper, and that paper's headline number is the most repeated figure in the field.
Where the term comes from
"GEO: Generative Engine Optimization", by Pranjal Aggarwal, Vishvak Murahari, Tanmay Rajpurohit, Ashwin Kalyan, Karthik Narasimhan and Ameet Deshpande, was first posted in November 2023 and accepted to KDD 2024. It introduced GEO-bench, a benchmark of user queries across several domains with web sources to answer them, and reported that GEO "can boost visibility by up to 40% in generative engine responses". The paper.
The same aim is also called answer engine optimization (AEO), LLM optimization and AI search optimization.
"Up to 40%" is 7.9 percentage points.
The authors released their benchmark results, including a baseline where each source was left unchanged. Each answer drew on five sources, so a source with no advantage would take about 20% of it. Unchanged sources took 19.4%.
The best method, adding quotations, took 27.3%. That is 7.9 percentage points, or 40.6% relative to the baseline, which is where "up to 40%" comes from. The same method made the result worse on 28% of queries.
Source. The GEO-bench leaderboard released with the paper · 989 queries with results for every method, 110 empty rows excluded · share of the generated answer as the authors' own measure scores it · the answering engine is not labelled in the file · re-derived by Vector Index.
How the top methods were produced
The code that rewrote the sources is public. Three of the four highest-scoring methods were instructions to fabricate. Adding quotations: "Add more quotes in the source, even though fake and artificial." Citing sources: "You may invent these sources but ensure they sound plausible". Adding statistics: "Add positive, compelling statistics (even if hypothetical)". The released code.
We do not recommend any of them. A business that published invented quotes, sources or statistics would be making false claims about itself, whatever an answer engine did with them.
What GEO work looks like without the shortcuts
For Google's AI features, Google says: "There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary." Google's documentation.
The rest is ordinary and checkable. Find out what each engine says about a business for the questions its buyers ask. Separate being mentioned, being cited and being recommended. Correct the facts and pages the answers rely on. Then ask the same questions again on the same engines. In our own review of the published research, we found no controlled before-and-after test on a live engine, so every recommendation should trace back to a stored answer rather than to a borrowed percentage.
AEO vs SEO vs GEO · GEO and AEO services · How we measure · Our scope and prices
Questions we get asked about this.
What does GEO stand for?
Generative engine optimization: improving how a business appears in answers written by generative AI systems such as ChatGPT, Gemini, Claude, Perplexity and Google AI Overviews.
Who coined generative engine optimization?
The term comes from the paper "GEO: Generative Engine Optimization" by Pranjal Aggarwal, Vishvak Murahari, Tanmay Rajpurohit, Ashwin Kalyan, Karthik Narasimhan and Ameet Deshpande, first posted in November 2023 and accepted to KDD 2024.
Does GEO really increase visibility by 40%?
In the paper's released benchmark results, the best method raised a source's share of the answer from 19.4% to 27.3%, a 40.6% relative gain, and made it worse on 28% of queries. It was measured on the paper's own benchmark.
What is the difference between GEO and AEO?
Both describe the same aim. GEO comes from the research paper, and AEO names the same work after the answer engines. Neither is a formal standard.
Is GEO replacing SEO?
Google says there are no additional requirements to appear in AI Overviews or AI Mode. A page still has to be indexed and eligible to show in Search with a snippet, so search fundamentals still apply.
We publish what we find.
We're measuring AI answers in Cincinnati and will publish what we find, including our own results, including when they're unflattering. No pitch attached.
One email a month at most, and you can leave with one click. We do not sell your address and we will not add you to anything else.