top of page
Start a conversation

ThinkWicker

The AI Shortlist Is Deciding Who Gets Considered

Writer: Owen Birch
Owen Birch
11 minutes ago
4 min read

More buyers now ask an AI assistant for options before they search. Here is the AI shortlist, why it matters, and how to see where your brand stands on it.


Ask an AI assistant for a good accounting firm for a small nonprofit, a pediatric clinic that takes your insurance, or a project management tool for a construction crew, and it will not hand you a page of ten links. It will list a few options, describe each in a sentence or two, and often suggest which one fits best.


That answer is the AI shortlist: the small set of brands an AI assistant names, describes or rules out when someone asks it for a recommendation. Brands on the list get compared. Brands left off it often never get the chance. Brands described inaccurately can be eliminated before the buyer ever visits their website.



Why the AI shortlist matters now


The clearest recent data comes from a survey Talker Research fielded in September 2026 among 2,000 U.S. adults who had used an AI tool for personal reasons in the previous three months. Among them:


  • 70% had discovered a brand, product or service for the first time through AI.

  • 64% had bought or used something from a brand they first found through AI.

  • 68% had ruled out a brand they were considering after learning something through AI.


When the assistant recommended something specific, few people ignored it. Of those who received a specific recommendation, 29% followed it completely, 24% followed it in part, and 28% researched further and then chose it anyway.


AI has not replaced the other ways people decide. In the same survey, search results (81%) and customer reviews (79%) were still rated as more influential than AI recommendations (73%). What has changed is the order. For a growing share of buyers, the assistant comes first and shapes which brands make it to the search.


One caution about the source: the survey was commissioned by We are Talker, a firm that sells AI visibility services to brands. Its figures are best read as directional rather than definitive.


By the time a buyer reaches the search bar, the AI shortlist has already decided which brands are still in the running.


Where it shows up


Among respondents whose decisions AI had changed, the most common categories were shopping (27%), food and restaurants (24%), health and wellness (24%) and financial decisions (22%). Those are not niche uses. They include the decisions behind choosing a clinic, a financial advisor or a place to eat, where trust matters and where a single sentence of description can tip the choice.


Consider a parent who asks an assistant which nearby clinics see children on weekends and accept their insurance. If a health center's website lists Saturday hours on one page and old hours on another, the assistant may describe it either way or leave it out of the answer entirely.


The parent will never see the correct page, because they never reached the website. The same is true for a contractor whose project list hasn't been updated in two years, or a law firm whose practice areas changed after a merger. The shortlist reflects whatever is written down, accurate or not.



The most common misconception


The most common misconception is that this is search engine optimization under a new name. There is overlap: both reward clear, accurate information about a brand. But a search results page shows the buyer several options and leaves the comparison to them. An assistant often does the comparing for them and presents a conclusion. The elimination happens before any website loads.


A second misconception is that a place on the list can be bought. A Harris Poll survey conducted for Quad in February 2026 found that 75% of Americans would trust AI shopping agents less if brand payments influenced their recommendations, and the same share would trust the brands less. Whatever paid options emerge, buyers have already said how they will read them. The durable path is to be described accurately in the places assistants draw on: the brand's own website, reviews and independent coverage.


Being left off the list is one risk. Being on it and described wrong is the bigger one.


A first application: ask it yourself


The simplest way to see your position is to look at it the same way a buyer would.


  1. Write five questions a real buyer asks before choosing an organization like yours. Use their words, not yours: "Which clinic near me takes Medicaid and has evening hours?" rather than "best FQHC services."

  2. Ask each question in two or three AI assistants. Use a browser where you aren't signed in, so personal history doesn't shape the answer.

  3. Record three things for each answer: whether your brand is named, how it is described, and anything stated about it that is wrong or out of date.

  4. Trace each error to its source. An outdated service list on your website, a gap in reviews or a missing or outdated profile on a listing site. Fix the source, not the symptom.

  5. Repeat every quarter. Answers change as your information changes.


Most organizations find at least one surprise in the first round. That surprise is usually the most useful thing they learn.



Where to go next


The AI shortlist connects to ideas ThinkWicker has explored in more depth. Discovery Without Search examines how people discover brands they didn't know to look for. Reputation Has Two Dimensions separates how well a brand is regarded from how widely it is known, which is the gap the shortlist exposes. Ingredient Legibility argues that what a brand clearly states about itself is what an audience believes, and an assistant is one more reader of that information.


Every AI assistant is one more reader of everything written about your brand. The first step is finding out what it has read.


References



Some ideas are worth discussing in the context of your organization.



 
 

Need more than insights?

bottom of page