The shortlist is the new page one.

Like most of us looking for a solution, we began that search on Google Search over the last 25 years. We type in a few words (we all got really good at this) and ultimately would click on a blue link, visit a website, and end up in a marketing funnel.

But since late 2022, something has changed, and continues to propagate. For many of us, across multiple demographics and verticals, we've migrated to a more buyer-friendly solution-finding experience, AI search.

For business owners and marketing teams, this shows up as a decline in organic search traffic, and it is happening to nearly everyone. Demand for most products hasn't gone anywhere. People are still searching. The top of the funnel has just moved off your website and into the AI answer.

And inside that answer, buyers search differently. They describe their situation in detail, explain their use case, and ask for the product most likely to fit. What comes back is a shortlist of a few names, not ten blue links. That shortlist is the new page one.

Naturally, the first thing most of us asked was, how does an LLM model decide? Where does the list come from? The short answer is, from us. From everything we've published over the years, the reviews and media we've earned, and everything others have written about us. The AI learned from it in training, and when a buyer asks a question today, it often improves its response by what it can find from a web search today.

The part we control most is our own website, and for a lot of us, the site has been telling the same human-centric story since the pre-AI search. Now, there is another type of visitor where we deliberately need to reduce the inference needed, say plainly what we do, who it's for, present real data, and back claims up with proof. From there the story spreads, to profiles and directories, to reviews, to comparison pages and forums, and eventually the press. The AI leans on what it sees repeated across sources it trusts.

Surprisingly, for well known big brands, it's not a shoo-in that they make the shortlist. Sure, it helps, but it's not a guarantee. A bigger/older brand, with more domain authority and backlinks, a larger ad budget, none of it guarantees a spot on the list. Each shortlist is an answer to one specific question, and the AI search model is going to pick the companies whose story (and proof) fits that question. We have seen well-known brands get left off answers that smaller competitors show up in, simply because the smaller competitor said it clearly, executed well on the details, and others repeated it.

Which brings us to the question that keeps us up at night. Are we on the buyer's shortlist or not? Back in the olden days (a few years ago), pre-AI search era, this was easy to answer. We tracked some keywords, got a position and a decent guess at search volume, and you knew where your brand stood. AI search doesn't work that way. There is no single static list to pull up.

Buyers ask hundreds of versions of the same question, conversing with AI in their own words, with their personalized chat history, across several models, and each answer is its own shortlist. Ask one prompt and we learn almost nothing. However, when you ask a few hundred, grouped by topic, a pattern shows up. Some brands keep reappearing. Those are the brands making the shortlists, the new page one.

So where your brand stands is a mention share, not a classic search engine ranking. It's how often your brand gets named across a topic, compared to how often your competitors get named. And since the same prompt can come back different under endless scenarios, and each AI model answers in its own way, that share has to come from asking many times, in each AI search system, over time.

Example data.

How often do we need to check AI search results to understand how visible our brand is in AI search for the topics we care about? Not every change in AI search answers means something. Some of it is just the AI being the AI. Some of it is real. A new review landed, more news coverage, a competitor published a page, the model got an update.

Telling those two apart comes down to how many AI search answers sit behind each reading. We can get there with breadth (more prompts per topic) or with frequency (running the same prompts more often). Either way, a share built on a few hundred answers holds steady week to week, and when it does move, we can trust that it moved.

Each one tells us something different. Running the same 6 prompts every day shows how those 6 exact questions are moving. Running 40 different prompts under one topic once a week shows how the whole topic is moving, because those 40 cover more of the many ways different buyers ask about the same need. Across 8 AI search systems, that's 320 answers behind one weekly reading of the topic.

In Hi, Moose AI Visibility tracking, you can choose a custom schedule (multiple times per day, once daily, weekly, weekdays, etc.) that matches the needs of your business.