How is the AVI Score calculated?

It is 60% Outcome and 40% Readiness. Outcome is measured by putting real buyer questions to ChatGPT, Gemini, Claude and Perplexity, sampling each question multiple times, and counting how often the brand is mentioned, recommended and cited versus its competitors — with confidence intervals. Readiness is measured by crawling the site for the signals AI systems rely on. The two combine into one 0–100 number.

Is the AI Visibility Index a tool?

No — that distinction is the point. A tool measures you privately. The index measures a whole field and publishes the ranking, so the number means something outside your own account. You can run the free Readiness scan on yourself, but the index itself is the published standard your category is measured in.

Which AI assistants does it measure?

Currently ChatGPT, Gemini, Claude and Perplexity — measured across all of them rather than any single one, because a score that only held inside one product would not be a standard. The index is built to add assistants as coverage expands, and each reading records the ones it ran on.

Can a brand pay to rank higher?

No. The index is vendor-neutral: no brand can buy a place, pay to move up, or have a competitor removed. A subscription pays for measurement and a remediation plan; it does not buy a position. The ranking would be worthless if it could be bought.

Is the AVI Score a guarantee that AI will recommend me?

No. It is a repeatable measurement, not a promise. AI answers are non-deterministic and change over time, so the score is a rate with a confidence interval, not a fixed fact — and improving it improves your odds of being named, not a certainty of it.

What it is

What is the AI Visibility Index?

The AI Visibility Index is the standard of record for how AI sees a brand. It publishes a score, out of 100, for how often AI assistants actually name and recommend a brand — measured across ChatGPT, Gemini, Claude and Perplexity — and ranks every brand in a category against the others. It is not a tool you run on yourself: it is the index your market is measured in.

Why it exists

For twenty years, being found was measurable. You had rankings, traffic, a dashboard you could open. Now a customer asks an assistant “what’s the best in this category?” and acts on the three names it returns — and when yours isn’t one of them, nothing tells you. There is no dip to trace, no alert, no Search Console for AI. The sale simply goes to whoever was named.

That blind spot is the reason the index exists. If an unmeasured channel is deciding who buyers choose, then measuring it — consistently, in the open, for a whole field at once — is the thing worth doing. An index is how a market makes something measurable.

What it measures

The AVI Score combines two halves that answer two different questions.

Outcome · 60%

Does AI actually recommend you?

Real buyer questions are put to the major assistants and sampled repeatedly. We count how often the brand is mentioned, recommended and cited — versus its competitors, with confidence intervals. This is the result that matters.

Readiness · 40%

Is your site built to be found?

The site is crawled for the signals AI systems rely on — crawlability, structured data, entity clarity, content depth. It is the fixable groundwork that moves the Outcome over time.

Weighted 60/40, they produce one number from 0 to 100. The full method — every question, every cohort, every interval — is on the methodology page.

What makes it an index

Plenty of products will scan your site and hand you a number. Four things make this a standard rather than a scan.

A published score: The AVI Score is a single number from 0 to 100, computed the same way for every brand measured, and published rather than kept in a private dashboard. A number that means the same thing for everyone, in the open, is what lets it be cited — by a brand, a journalist, or an assistant itself.
An open method: The questions asked, the competitors a brand is measured against, the sampling, and the confidence intervals are all stated. A standard whose method is hidden is just an opinion with a number attached. Ours is on the methodology page, and it is versioned, so a change to how we measure is never mistaken for a change in a brand.
The whole field: An index ranks everyone in a category, not only the brand that asked to be scored. That is the difference between a reading and a ranking: you can see where you stand against the businesses a buyer is actually choosing between, because they are measured too.
Vendor-neutral: The index is a referee, not a player. No brand can buy a place in it, pay to move up it, or have a rival removed. Its value depends entirely on that being true, so it is not for sale.

How to read an AVI Score

A score is a position on one scale, from largely invisible to reliably recommended. The bands name where a brand sits.

Score	Band	What it means
90–100	AI Visibility Leader	AI reliably names and recommends the brand across assistants.
80–89	Advanced	Strong, consistent presence with a clear technical foundation.
60–79	Established	Regularly surfaced, with room to widen the lead.
40–59	Developing	Present but inconsistent — visible on some questions, absent on others.
0–39	Emerging	Largely invisible in AI answers, whatever the quality of the business.

How it differs from a monitoring tool

A monitoring tool tracks your brand in a private dashboard, for you, on a schedule. That is useful, and it is not what an index is. The index publishes one comparable number for a whole field, so where you stand is a public fact rather than a private chart — the way a credit score or a market index is a reference everyone shares, not a report only you can see. If you want daily internal alerting, a monitoring tool fits better; if you want a defensible external number and to know where you rank against the field, that is the index.

The honest limits

It is a measurement, not a guarantee. Improving your score improves your odds of being named, not a certainty of it.
AI answers are non-deterministic. Ask twice and the wording, and sometimes the brands, differ. That is why the score is a rate with a confidence interval, built from many samples rather than one.
It is built category by category. A field is only an index once its cohort is measured, so coverage grows one category at a time rather than claiming the whole market at once.
It measures the assistants named, not all of them. The index covers ChatGPT, Gemini, Claude and Perplexity today, and every reading records which of them it ran on — so a score published before an assistant was added does not claim it. Being named by one says little about the rest, which is exactly why the index measures across several.

Common questions

What is the AI Visibility Index?: It is the standard of record for how visible a brand is to AI assistants. It publishes a score, out of 100 — the AVI Score — combining how often AI actually names and recommends the brand (Outcome, 60%) with how well its website is built to be found and cited (Readiness, 40%), and it ranks every brand in a category against the others.
How is the AVI Score calculated?: It is 60% Outcome and 40% Readiness. Outcome is measured by putting real buyer questions to ChatGPT, Gemini, Claude and Perplexity, sampling each question multiple times, and counting how often the brand is mentioned, recommended and cited versus its competitors — with confidence intervals. Readiness is measured by crawling the site for the signals AI systems rely on. The two combine into one 0–100 number.
Is the AI Visibility Index a tool?: No — that distinction is the point. A tool measures you privately. The index measures a whole field and publishes the ranking, so the number means something outside your own account. You can run the free Readiness scan on yourself, but the index itself is the published standard your category is measured in.
Which AI assistants does it measure?: Currently ChatGPT, Gemini, Claude and Perplexity — measured across all of them rather than any single one, because a score that only held inside one product would not be a standard. The index is built to add assistants as coverage expands, and each reading records the ones it ran on.
Can a brand pay to rank higher?: No. The index is vendor-neutral: no brand can buy a place, pay to move up, or have a competitor removed. A subscription pays for measurement and a remediation plan; it does not buy a position. The ranking would be worthless if it could be bought.
Is the AVI Score a guarantee that AI will recommend me?: No. It is a repeatable measurement, not a promise. AI answers are non-deterministic and change over time, so the score is a rate with a confidence interval, not a fixed fact — and improving it improves your odds of being named, not a certainty of it.

Go deeper

See where you stand in the index

Start with the free Readiness scan — the readable half of your AVI Score, in seconds, no signup.