Models Operate on Topics, Not Keywords

More players in the AI search industry are catching up to what Bluefish has been saying all along: Traditional SEO strategies can only take a brand so far. In order to get noticed and recommended by AI, brands need to become authorities in the topics most relevant to their customers. This logic should inform how marketers measure their brand’s AI visibility.  

A recent study commissioned by Semrush, an Adobe-owned SEO tool, found that strong SEO fundamentals aren’t enough for AI systems to consider a brand to be a leader in its category. Nearly half of the brands that Adobe says are category leaders on ChatGPT were found to have worse SEO metrics than the second-most mentioned brand.  

“Strong SEO helps you show up,” Adobe said, “but once several brands are in the running, deep coverage of the specific topic is what turns you from runner-up to owner.” 

Microsoft is also emphasizing the importance of topics in AI search. In June, the company updated its AI visibility reporting tools with a new feature called Topics, which groups together AI citations that share a common theme. 

Microsoft states the premise plainly: 

"AI systems reason across concepts and themes rather than isolated keywords. Topics help publishers understand visibility in the same thematic structure that modern AI systems use to organize information.” 

How Topics work  

Topics are the factors that people use to assess products, such as comparing horsepower, comfort, and milage when buying a new car. Unlike traditional search engines, which surface webpages with matching keywords, AI platforms like ChatGPT, Claude, and Google Gemini understand the intent behind a prompt, and combine that intent with the information they already know about the consumer from prior conversations to generate deeper, more personalized responses.  

When an AI system is prompted, the model rewrites the prompt into a grounding query. Grounding queries are the searches AI performs to gather reliable information before generating an answer. Microsoft’s new tools group grounding queries that are thematically connected into larger Topics. The company says that Topics “can help publishers identify emerging areas of authority, discover gaps in topical coverage, and better understand how AI systems semantically group related content.”  

Why Topics matter 

Adobe and Microsoft’s embrace of Topics backs up a key part of our philosophy at Bluefish: AI visibility is earned through retrieving data across a theme rather than collected query by query. Publishers should work to optimize their content around a set of key themes rather than generating keyword-heavy pages that may rank well on traditional search, but are lacking in depth.   

This is why AEO players emphasizing "real" prompts mislead the market, or worse, simply don’t understand how AI models work. No platform exposes what consumers typed verbatim. What truly matters in optimization is extracting the right performance for the consumer intent that matters most to a brand, with data integrity across a large volume of prompts, and enough audience granularity to make it actionable. Bluefish sets that standard. 

Here are some takeaways from this recent guidance that brands can implement today: 

• AI interprets intent.  Systems like ChatGPT and Google AI Overviews respond to what a person is trying to accomplish instead of the exact words they type.  

• The audience changes the answer. Memory, conversation history, and context personalize every AI response. If a first-time buyer and a brand loyalist both give identical prompts to the same model, the model response will still be different depending on the user’s identity.  

• Citations are one of the clearest signals of AI influence. Brands earn citations through authoritative, high-quality content that AI systems deem relevant to a user’s intent. Knowing which sources drive citations, and how they shape AI, can help brands strengthen their content structure and refine their topical authority. 

When brands partner with Bluefish, they gain the ability to measure the dimensions LLMs operate on, including intent, audience, and citation impact across categories. To learn more about Bluefish and how we help brands win in AI, reach out to the team at bluefishai.com/contact/sales.