The next speaker at the SEASON 2026 conference in Hamburg is Hadeer ElMalah, whose focus is on both search engine optimisation (SEO) and generative engine optimisation (GEO). The algorithms that manage SEO on the search engine side have evolved greatly over the years, but are now further disrupted by AI-enhanced search processes. For marketers, the big question is now how to get brands cited in AI responses.
This requires a more advanced understanding of how generative AI engines choose what to cite. Very basically, retrieval-augmented generation relies on indexing documents; retrieving potentially relevant materials in response to a query; and generating a human-readable answer that builds on those materials – however, this is only a very simple representation of a much more complex picture which now also involves pre- and post-retrieval elements.
More advanced retrieval-augmented generation systems now focus not just on providing a set of the best single matches for a query, but also the best coverage of the various aspects of a query – covering multiple semantic clusters associated with a distinct query.
Hadeer explored this using a set of Websites associated with the SIXT car rental company: her strategy used an AI writing agent to optimise partner Website content for a multi-cluster context, while also ensuring that the AI-generated content remains accurate and reflects the actual vehicles available from SIXT, and making sure that language is confident. This increased citation in generative AI results from various platforms by a factor of three, compared to Websites were such changes were not implemented – and such improvements were sustained even after the experiment ended.
This shows that retrieval-augmented generation systems are susceptible to such GEO interventions, and this needs to be further – especially also as such systems evolve further.
