"For what feels like the first time, journalists and AI chatbots are on the same team: that of objective language. Large language models typically favor informational rather than promotional language, said Jack, "almost like a scientific research paper." With that in mind, one tip Unusual often shares with clients is to "give some kudos to their competitors" to highlight their own strengths by comparison."
"Generative engine optimization, or GEO, is starting to feel like an unwinnable game. Every time brands think they've figured out how it works, there's a new level to beat. The latest challenge, according to Will Jack, co-founder of AI discoverability startup Unusual, is less about showing up within AI search and more about standing out. Unusual, which announced a $3.6 million funding round earlier this month, analyzes how brands appear within AI search and provides suggestions on how to better rank."
"Instead, said Jack, marketers should prioritize earned media and create dedicated AI content that's designed to be "found by search engines and not humans." However, it's important for transparency's sake that humans can access this content if they want to, Jack added, which means no "cloaking." But SEO should "outrank" your dedicated AI content, he said, since the latter isn't meant to"
Generative engine optimization (GEO) has become increasingly complex as AI search systems evolve and change ranking criteria. Unusual analyzes how brands appear within AI search and offers recommendations to improve ranking, backed by recent funding and team expansion. Large language models favor informational, objective language over promotional claims, treating flowery marketing as unhelpful. Brands benefit from emphasizing earned media, offering comparative praise of competitors, and producing dedicated AI-targeted content designed to be indexed by search engines. That AI-targeted content should remain accessible to humans (no cloaking), while traditional SEO should still outrank specialized AI content.
Read at AdExchanger
Unable to calculate read time
Collection
[
|
...
]