GEO: The Science of Making Brands Visible in AI Answers
People increasingly find what they're looking for not in lists of blue links but in the answers produced by generative engines like ChatGPT, Perplexity, and Gemini. In response to this shift, academia has named a new discipline: GEO (Generative Engine Optimization). And crucially, content's visibility within those answers can be improved in a measurable way.
What the research found
In a pioneering paper presented at KDD 2024, Aggarwal and colleagues built an evaluation framework that measures how often content is cited in AI-generated answers. They then tested how targeted adjustments to content structure affected that visibility.
The result: with the right content adjustments, a source's likelihood of being surfaced in AI answers rises measurably. This challenges the assumption that "what the AI shows is a total black box"; brands can manage their position in next-generation search as active participants rather than passive bystanders.
What it means for brands
Classic SEO focuses on getting a user to click through to your page. GEO plays for a different goal: being named as a source inside the answer itself. When an AI responds to a question, it chooses which brands, products, and sources to build its answer on; being left out of that selection means being ignored in the new form of visibility.
This calls for rethinking how content is structured: an information architecture that gives clear answers, content that grounds claims in sources, structural markup that machines can also read, and a consistent presence that positions the brand as an authority on a topic. These help in both classic search and AI answers.
GEO sits at the center of how Scope thinks about digital visibility, because the future of search rests not on a single channel but on managing classic search and AI answers together.
In the end, visibility is no longer just the question "where do I rank on Google?" — it's "does the AI mention me?" GEO is the science-based answer to that question.
References
- Aggarwal, P., et al. (2024). GEO: Generative Engine Optimization. Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD '24). arXiv:2311.09735. https://arxiv.org/abs/2311.09735
