How to check if AI assistants recommend your brand

The buying journey moved, and most brands missed it

For two decades the question was "do we rank on the first page of search?" That question still matters, but a new one now sits in front of it. A growing share of buyers open an AI assistant and simply ask: "what's the best tool for X?" or "what should I use instead of [a competitor]?" The assistant answers with a short, confident shortlist of three to five names. If yours is on that list, you enter the consideration set effortlessly. If it is not, you were eliminated before the buyer ever ran a search, visited your site, or saw your ad.

This is a quieter and more total kind of invisibility than poor search ranking. With search, at least you appear on page two; a determined buyer can scroll. With an AI recommendation, the names that are not spoken simply do not exist for that buyer in that moment. And because the conversation is private, you have no log, no analytics, no impression count telling you it happened. You lose the deal and never know there was one.

So the first job is not optimization — it is simply finding out where you stand. Most founders have never actually asked the assistants their own buyers' questions and written down the answers. When they do, the result is usually a surprise.

The four things to measure

"Do they recommend me?" is too blunt a question. A useful audit breaks visibility into four dimensions, each scored on its own:

• Presence. When asked about your category, does any assistant name you at all? This is the floor. If presence is zero, nothing else matters yet.
• Position. When you are named, where do you land? First in the list carries far more weight than fifth. Moving from "mentioned in passing" to "named in the top three" is a different outcome entirely.
• Sentiment. How are you described? "A solid, well-regarded option" and "a newer entrant with limited reviews" are both mentions, but they send the buyer in opposite directions.
• Accuracy. Are the stated facts about you correct? Assistants sometimes attribute the wrong features, wrong pricing, or wrong category to a brand. An inaccurate recommendation can be worse than none, because it sets a false expectation you then have to unwind. To track all four dimensions over time, compare the best AI visibility tools.

Scoring each dimension separately turns a vague worry into a clear scoreboard. You stop asking "are we visible?" and start asking "our presence is fine but our position is weak and one assistant has our pricing wrong — fix those two things."

How to run the check

1. Write your buyers' real questions. Not "is [my brand] good" — buyers don't ask that. They ask "best [category] tool", "[category] software for [their situation]", and "alternatives to [a competitor]". Collect ten to twenty of these.
2. Ask across the major assistants. Coverage matters — different assistants draw on different sources and give different shortlists. Checking only one gives you a skewed read.
3. Record the four dimensions for each answer. Were you named? Where? How were you described? Was anything wrong? A simple grid turns gut feel into data.
4. Repeat on a schedule. A single snapshot is a starting line, not a trend. The number that matters is whether your visibility is rising or falling over weeks, especially after you make changes.

Doing all this by hand is slow, easy to do inconsistently, and tedious enough that almost nobody keeps it up. That is exactly why a dedicated visibility tool exists: it runs your prompts across assistants on a schedule, scores the four dimensions automatically, and shows you the trend rather than a one-off impression.

From measuring to acting

Knowing your score is only valuable if it tells you what to do next. The good news is that AI recommendations are not random — they are built from the public record about your category. Assistants lean on the same authoritative sources buyers trust: software directories, comparison and review pages, well-structured write-ups, and the broader web's consensus about who the credible options are. When that record is thin or stale for your brand, you are invisible. When it is rich, accurate, and citable, you give assistants a reason to name you.

That means the work is concrete and doable. The highest-leverage moves are usually: getting listed and reviewed in the directories assistants cite most, publishing accurate and clearly-structured material that answers your category's real questions, correcting the specific factual errors an audit surfaces, and steadily building the kind of public footprint that makes naming you the obvious choice. None of it guarantees an assistant will name you, but it is the part of the answer you can actually influence.

Where MentionFox fits

MentionFox asks your buyers' questions on a schedule, records whether your brand is named, where, and whether your own page is linked, and scores presence, position, sentiment, and accuracy into a single number out of 100 so you can see the trend. Its Win now board lists buyer questions where no single source owns the answer yet, with the domains actually cited in those answers, so you know where new material has a chance. It measures and points to openings; it does not change how an assistant answers. Agencies can run the same check for each client they manage.

Questions, answered