The top tier of ads remained relatively consistent across platforms, but the middle of the pack proved volatile. Ramp, for example, was ranked No. 3 in Claude, No. 5 in ChatGPT, and dropped to No. 11 in Gemini. Dove followed a similar pattern, placing No. 2 in Claude, No. 7 in ChatGPT, and No. 13 in Gemini.
In a landscape increasingly shaped by LLMs, ad memorability appears to be determined less by spend and more by how easily a campaign can be summarized, cited, and reinterpreted by machines.
Where–and why–LLMs disagree
The analysis highlights why AI models remember the Super Bowl differently.
ChatGPT, for example, favored ads that could be explained clearly in conversation, producing balanced, narrative-driven responses. Claude prioritized emotional resonance and purpose, elevating campaigns that emphasized values or storytelling.
Perplexity leaned on information and citations, giving weight to ads with strong press coverage or widely shared recaps. Gemini focused on mainstream visibility and tech relevance, rewarding campaigns with broad awareness or media presence.
Grok, meanwhile, was attuned to culture and sentiment, amplifying ads that sparked humor, controversy, or social reaction.
LLMs favor challenger brands
Several challenger brands punched above their weight in AI recall, per Emberos.
Ramp, which ranked No. 4 overall, benefited from the clarity of its B2B messaging, which made it easy for models to summarize and cite. It also appeared frequently in business and tech recaps, boosting its ranking in Claude and ChatGPT beyond what public popularity polls suggested.
Invest America, at No. 11, similarly over-indexed: Its finance-focused ads lent themselves to structured, informational narratives, giving them consistent visibility across prompts.
Even ai.com, at No. 18, earned a boost thanks to what Emberos calls “AI keyword gravity,” with models favoring its AI-adjacent branding—particularly in Gemini—while other platforms kept it mid-tier.
Emberos’ AI Influence Index suggests that several challenger brands enjoyed higher AI visibility than “general-popularity expectations, likely because their messaging was unusually easy for LLMs to summarize, cite, and re-contextualize in follow-on prompts.”
