TLTR: marketing companies and agencies work under improving site's AI visibility, which can bias data into junk'ish direction For example, a question to the ChatGPT interface what's better, Apple's new phone or Samsung's new phone, and the LLM is going to tell the reader that Apple is better because Samsung's new phone ain't no good, and the written material that produced that response was all paid for by Apple.
It can be influenced by what resources LLM gathers to give you an output, thing like grok can access twitter's materials, Gemini - data from google, etc
Models which are allowed to search the internet freely, if sources are not specified in prompt, will focus on sites which are optimized to be found by LLM.
It's a new "trend" in marketing community, before GPT era marketing agencies were obsessed with SEO, and now it's slowly transitions to AEO
Companies like SEMR and SMWB are into tools that, suppose, to improve visibility in AI search to make some money on hype, which in the end deteriorates LLM's outputs (Pay to Win) |
