As generative AI tools like ChatGPT and Google’s AI Overviews muscle into search territory, ad agencies are scrambling to make sure brands don’t get buried in digital obscurity. Shops like Jellyfish, Wpromote, and Kepler are spinning up dedicated AI search teams to help clients get a leg up with language models (LLMs), measuring how brands are represented, ranked, and, hopefully, recommended. Haleon, for example, is running tests on Meta’s Llama model to see how its Advil and Emergen-C products show up in LLMs’ answers—and making changes to influencer briefs based on what they find.
Jellyfish and Collectively rolled out a tool called Chorus, which evaluates the AI search impact of creator marketing, i.e., how those Instagram #ad posts actually shape what AI knows about a brand. Haleon calls the early learnings “illuminating,” if not quite enough to quantify results yet. Wpromote is pairing third-party tech with proprietary tools for Travelpro, which saw a 154% bump in AI search mentions after a pilot effort. And Kepler has at least 10 clients tinkering with AI perception strategies, using internal tech and LLM APIs to gauge—and tweak—how brands are understood.
Tactics now include tagging websites with llms.txt files, and breadcrumbs crafted specifically for the bots crawling AI training data. Even WordPress plugins (thanks, Yoast) are getting in on the act. As one exec put it, LLMs are now customer zero—gatekeepers sitting between a brand and the purchase path. It’s no longer enough to rank on Google – you’ve also gotta charm the robot answering everyone’s first question.
Full story at Digiday.