Transcript:

Hey, I’m Chris Long with Nectiv, and in today’s GEO tip, I want to talk about how marketers are going to need to better connect their brand name or brand entities to the individual queries or prompts that users might be searching in tools like ChatGPT or Google’s AI mode.

So, this kind of ties back to how large language models work. They are essentially predictive engines, right? They guess the word that is statistically most likely to come next based on their training data. So if you enter a given prompt, they will embed that prompt and then, based on their embeddings—their worldview—they will start to predict the word that’s most likely to come next.

For example, if you ask, “Hey, what’s the best e-signature software?” they’re going to have a list of brands that, based on their training data and based on their embedding model, are statistically most likely to appear next after that. So what you want to do is try to pair your brand with a response and get it associated with a word, term, or phrase in that neighborhood.

Right? So this is a very simplistic example here. I did a search in Google’s AI mode for “one of the best luxury pergolas.” I can see some of the sites getting sourced here — and StruXure is one that keeps coming up over and over when I perform these types of searches.

And what I can find is they actually have a luxury pergola page that gets sourced here. And when you end up clicking that, I like to pay attention to the highlighted text, because it might give some insight into what passage of text was used to train the model.

We can see that on their site they say, “The most trusted choice for homeowners, businesses, and builders.” They say, “StruXure pergolas are a top choice for business and builders.” So they’re directly connecting their brand and product to the term “best” — right in their own content on an optimized page they already have for the term “luxury pergolas.”

This is a good example of connecting a brand to a prompt or query that someone might search in an AI model.

And I know this is pretty marketing heavy, but I think this is something a lot of brands would benefit from — going through your pages (your product pages, your “how to choose” pages, your buying guides) and adding language that directly connects your brand to the target query or prompt you’re optimizing for.

That way, when a model trains on it, it might be more likely to associate your brand with the actual prompt a user inputs.