The question used to be "does my page rank." Now it is "does the assistant say my name."
When a buyer asks an AI assistant which tools to look at, the answer is a short list of brands, sometimes with a few links underneath. If you are not in that list, there is no page two to scroll to.
You can check where you stand in about ten minutes, by hand, with nothing but a spreadsheet. This post shows you how. It also shows why a single check tells you less than you think, and how to turn what you find into work someone owns.
Pick the questions your buyers actually ask
Your check is only as good as the questions you ask.
Start with 10 to 12 questions written the way a buyer would type them, not the way your website describes you. Pull them from sales calls, support emails and the search terms already in your reports. A good set covers a few kinds of question:
- Category questions: "What are the best tools for [what you do] for a small B2B team?"
- Comparison questions: "[Competitor] vs [you]" or "alternatives to [competitor]."
- Problem questions: "How do I [the job your product does] without [the pain it removes]?"
Leave your brand name out of most of them. Asking "is [your brand] any good" tells you how the assistant describes you once you are named. It does not tell you whether a buyer who has never heard of you would find you.
Write the list down and keep it. You will ask the same questions again next week, and the comparison only means something if the questions stay the same.
Ask, then write down what you hear
One answer is an anecdote. Five answers to the same question start to show a pattern.
Ask each question in the assistants your buyers use. For most B2B teams that means ChatGPT, Perplexity and the AI Overview at the top of a Google search. A few habits keep the check honest:
- Use a fresh session or a private window, so your own history does not shape the answer.
- Ask each question more than once. The same question can return a different list of brands a minute later.
- Turn on web search where the assistant offers it, since that is how most buyers will see it.
For every answer, record the date, the assistant, the question, every brand named, whether you were one of them, and any links the answer cited. A spreadsheet with one row per answer is enough.
Read the answers for what they say about you
Being mentioned and being recommended are two different results.
Go back through your rows and mark four things for each answer:
- Named: your brand appears anywhere in the answer.
- Recommended: the assistant suggests you as a good fit, not only lists you.
- Cited: one of the links under the answer points to your site.
- Described correctly: the assistant puts you in the right category and gets what you do right.
Each one points somewhere different. Named but never recommended usually means the assistant knows you exist but has little to say in your favor. Recommended but never cited means other sites are doing the talking for you. Described wrong is the one to fix first, because every answer that repeats the mistake sends the wrong buyer your way, or sends the right one elsewhere.
When a competitor shows up and you don't
The links under a competitor's answer are your to-do list.
When a competitor is named and you are not, open what the assistant cited for that answer. You will usually find one of a few things: a comparison page on the competitor's own site, a review or listicle on a third-party site, a forum thread, or a help page that answers the question directly.
Each of those becomes a specific task. "Write a page that answers [the exact question] in the first paragraph." "Ask to be added to [the listicle that keeps getting cited]." "Correct the category on [the profile the assistant is reading]." Give each task an owner and a date. A gap that lives only in a spreadsheet is still a gap next month.
Check again next week, and the week after
AI answers move. A check you run once is a snapshot of a moving target.
Assistants change which sources they read, competitors publish new pages, and the same question can return a different list from one week to the next. Run the same questions on a schedule, weekly if AI search matters to your pipeline, monthly if it is one channel among many, and compare against your last run rather than against your hopes.
This is one check among several. Your website needs checking for broken pages and slow loads, your paid search for spend going to the wrong terms, your cold email for deliverability before the next send. Each of those is the same loop: check on a schedule, find what changed, turn it into a task someone owns, and check again to see if the fix worked.
In Launch Point, that loop is a routine. The AI search visibility routine asks your buyer questions in each assistant on a schedule, shows who gets named and cited instead of you, and turns each gap into a task for you to approve. It sits beside the routines for your website, paid search and cold email, so AI search becomes part of the work instead of a report you read once.
Where to start
Start with one honest look at where you stand today.
Run the manual check above with your top 10 questions this week. If you want a faster first look, the free AEO report at report.launchpointbrands.com asks your buyer questions in ChatGPT, Perplexity and Google AI Overviews and shows who gets named instead of you.
When you are ready to make it a check that runs itself and lands as tasks in front of your team, book a demo and we will walk you through it.
