Measure and improve · 6 min read
How to Read Your AI Visibility Report
A report has a lot of numbers on it. Here is what each one means, which ones to look at first, and which to stop worrying about.
The short answer
A Found in AI report records, for every tracked question and each of five AI tools, whether you were named, whether your site was linked, which competitors were named and your position among them. It rolls that into a 0 to 100 score and grade, a trend, a leaderboard showing the share of answers naming each business, what changed since your last scan, and a list of gaps.
Named and linked are different things
Being named means the answer mentions your business anywhere in its text. Being linked means it also links to your website or cites it as a source. You can be named without being linked, for example when a tool mentions you based on a review site and links to that site instead of yours.
Both matter for different reasons. A mention puts your name in front of the customer. A link to your own site gives them a direct path to you and shows the tool read your pages. If you are often named but rarely linked, the tools are probably learning about you from other sites, which points at your own pages as the thing to strengthen.
Position: where you sit in the answer
When an answer names several businesses, the report records where you appear among them: first of three, third of four. Position gives texture to a simple yes or no. Being named fifth in a list of five is still being named, but it is a different result from leading the answer.
Do not over-read a single position. The order in an AI answer can shift from one asking to the next. Look at your typical position across many answers.
The score, the grade and the trend
The AI Visibility Score, from 0 to 100, and its letter grade summarise how often you were named and linked across all your tracked answers. It is useful as a single headline number and for comparing one scan with another.
The trend is more informative than any single score. AI answers vary from scan to scan, so a few points up or down between two scans is often noise. A steady direction over several scans is the signal. Read the score the way you would read a monthly sales figure: compare it with its own history, not with a number you hoped for.
The leaderboard: who the tools reach for
Across all your tracked answers, the leaderboard shows the share that name each business, you included. It answers the question 'when customers ask about my trade, which businesses do these tools keep mentioning?' Competitors appear the way the answers name them.
This is often the most eye-opening part of a first report. The business that dominates may not be the one you consider your main rival. It may be a directory, a national chain or a competitor with a strong review profile.
The per-tool view and 'since your last scan'
Break the results down by tool. It is common to do well in one and poorly in another. A pattern by tool can point at a cause: weak results only in Google AI Overviews suggests looking at your Google presence, weak results only in ChatGPT suggests checking crawler access and your Bing coverage.
Every report opens with what changed since your last scan: questions where you are newly named, and questions where a competitor took your place. Start there each time. It is the fastest way to see whether anything you did recently coincides with a change.
- First look: what changed since your last scan
- Then: the trend line, not the single score
- Then: the leaderboard, to see who is being named instead
- Then: results by tool, to spot patterns
- Finally: the gap list, to decide what to do next
Your checklist
- Open each new report on 'since your last scan' before anything else
- Compare your score against its own history, not a target number
- Count how often you are named without being linked
- Note the top three businesses on the leaderboard and whether they surprise you
- Look for any tool where your results differ sharply from the rest
- Pick one or two gaps to act on before the next scan
Common questions
Why did my score drop when I did not change anything?
AI tools change their models and sources, and competitors publish. Answers to the same question can differ between scans. Look at the trend over several scans before drawing conclusions.
Does a high score mean customers will find me?
No. It is a measurement of what the tools said to your tracked questions on the day of the scan. It is useful for tracking direction, not a promise about any particular customer.
What should I focus on first?
The gaps. They are the specific questions where a competitor was named and you were not, and each comes with a suggestion of what to publish.
Read next
Choosing Which Questions to Track
Tracking the wrong questions gives you a precise measurement of something that does not matter. Spend real time on the list.
Read the guide →Turning Visibility Gaps Into Content
The gap list is the most practical part of your report: a set of specific questions, the businesses named instead of you, and a suggestion for each. Here is how to work through it.
Read the guide →Tracking Competitors in AI Answers
Your competitors in AI answers are whoever the tools name for your customers' questions. That list is worth studying closely, because it is often not the one in your head.
Read the guide →See where you stand today.
Found in AI asks five AI tools the questions your customers ask and shows where you are named, where you are linked and where a competitor is named instead.