AI search / query fan-out

Yesterday I spoke about how AI search is already more important than traditional search. Today I wanna talk about the difference and what it means for your business and website.

When I say AI search, I'm referring to anytime someone's using a large language model or Google's AI search results to find an answer. And if you have a business that sells an actual product or service, the only answers you really care about are recommendations.

You want the AI to recommend your brand as a solution. While there are some differences between each AI model and platform, they fundamentally work the same.

When a user goes to one of these AI models and asks for a recommendation, that model is almost always going to perform a search, or rather a series of searches. Going to evaluate not just the prompt you entered, but also everything it knows about you from previous conversations and this conversation.

It's going to determine what information it needs to fully address your question. And then it's typically going to run three or more searches.

This is called query fan out...

ntal ways in which AI search is different than classic search. First of all, the queries that it searches for are way more specific and personalized than any query a human would typically type into Google.

You're no longer competing over a handful of super competitive search terms. The strategy shifts to creating lots of hyper personalized content targeting specific use cases demographics.

The second is that humans will typically only look at the top three results, and they're much more likely to click through to position number one. AI, on the other hand, will commonly look through ten or twenty results, and they don't seem to have any bias towards websites that are ranking higher in those results.

So instead of fighting over position one, as long as you're in the top twenty results, you're in an equal playing field with everyone else in those results. This lowers that competitive bar even further, and it allows you to target keywords that otherwise wouldn't have made sense.

And the third way that's different, you don't have to focus on getting your website into those results...

. It's important that you're tracking which pages the large language models are citing before they recommend you or your competitors.

Surgically placing recommendations for your brand on those pages can be very effective. Tomorrow I'm gonna be talking about a new Google patent that has a potential to replace websites entirely.