What is an AI brand mention worth?
Short answer: nobody can quote you a price, and the figures that circulate are estimates built from somebody else's funnel. A mention's value is derived, not looked up. This page gives you the derivation.
The answer in one line
One mention is worth consideration lift × close rate × customer value. Everything else on this page explains where those three numbers come from, and why the first one is the only hard part.
Why no one can quote you a price
Every other channel has a price because it has a marketplace. AI answers have neither, and three structural facts keep it that way.
A search result is a published page a crawler can find. An AI answer is generated for one person, for one request, and is never published anywhere. There is no inventory record for a pricing model to attach to.
Search and display ads have a clearing price because advertisers bid against each other for a defined slot. No model currently sells a position inside a generated answer, so no price forms.
A mention without a link produces no referrer. The buyer reads your name, remembers it, and searches for you directly days later — where your analytics records a branded visit and credits the mention with nothing.
Treat any published per-mention dollar figure as a marketing artefact. It can only have been produced by modelling one company's funnel, which makes it a worked example of the method below rather than a market rate.
The formula, and where each number comes from
Monthly value is V × M × C × R × L. The first two terms size the opportunity; the last three price it. Which is why a single mention is worth C × R × L on its own.
| Term | What it asks | Where you get it |
|---|---|---|
| V | Prompt volume How many times a month is the question asked? | Start from the questions your buyers actually ask, not keywords. Your sales calls, support tickets and onsite search are better sources than a keyword tool, because AI prompts are written as full questions and phrased the way people speak. |
| M | Mention rate What share of those answers name you? | This is the one term you can measure directly. Run the prompt set against each model on a schedule and count the answers that name the brand. It is the metric this platform reports as brand mention rate. Brand mention rate → |
| C | Consideration lift For what share of readers does being named change the shortlist? | The hard term, and the one no dashboard reports. Approximate it with your word-of-mouth or referral conversion rate: an AI answer behaves like a recommendation from a trusted third party, not like an ad impression. |
| R | Close rate What share of considered buyers become customers? | You already have this. It is the same rate your CRM reports for the rest of the funnel. |
| L | Customer value What is a customer worth? | Average deal value for one-off purchases, lifetime value for subscriptions. Use the number your finance team already defends, not a more flattering one. |
The term everyone gets stuck on
Consideration lift is not on any dashboard, so most teams either skip the exercise or invent a number. Three workable approximations: use your existing referral conversion rate, since an AI answer functions as a third-party recommendation rather than an ad; add a "where did you first hear about us" field at signup and read it after a quarter; or watch whether branded search volume moves with mention rate when nothing else in the mix changed. None is precise. All three are better than a figure copied from a vendor's blog post.
Why two mentions are rarely worth the same
Mention rate counts. It does not weigh. These five differences separate a mention that moves a deal from one that changes nothing.
Position inside the answer
Named first, in a sentence of its own, is not the same as ninth in a list. Readers of a generated answer behave like readers of any ranked list: attention concentrates at the top and decays fast.
Cited or only named
A mention that carries a link to your page can send traffic and can be attributed. A mention with no link produces no referrer at all — it still moves the buyer, but it will never appear in your analytics.
How you are described
Being named alongside "expensive and hard to implement" is worth less than the same position alongside "the option most teams start with". Presence and perception are different measurements, and only one of them is a count.
AI Perception ScoreWhat the prompt was for
A mention inside "best tool for X" sits at the decision. The same mention inside "what is X" sits at the start of the research. Both have value; they are not the same value, and averaging them hides the difference.
Which model produced it
A mention is only worth what the audience seeing it is worth. Track the models your buyers actually use rather than treating eight models as one number.
Supported AI modelsThe other half of the question: what does one cost?
People asking for the price of an AI brand mention usually want to know what it takes to get one. You cannot buy it, so the cost is the work: the content, the original data and the third-party coverage that put your brand into the sources the models draw on. Divide that spend by the mentions it produced and you have a cost per mention for your brand — the only price that actually exists.
Where the economics differ from advertising
Paid media stops the day the budget stops. An earned mention does not: once the source exists and the model draws on it, the mention repeats every time the question is asked, at no additional cost, until the source or the model changes. That makes the spend look expensive against a first-month CPA and cheap against a second-year one — which is the reason it is usually measured on the wrong horizon.
Measuring the terms you can actually measure
Of the five terms, two are observable from outside your business: how often the question gets asked, and how often the answer names you. Both require running a fixed prompt set against the models on a schedule, because there is no archive to look them up in afterwards. That is what this platform does.
Common questions
What is the value of a brand mention in AI?
There is no published price, and any single dollar figure you see quoted is an estimate someone built from their own funnel. The value of one mention is the consideration lift it produces, multiplied by your close rate, multiplied by what a customer is worth: C × R × L. The first term is the hard one, and it is specific to your category and your buyer.
What is the price of a brand mention in AI?
Nothing, because there is nothing to buy. No AI model currently sells placement inside a generated answer, so there is no auction, no clearing price and no rate card. What has a price is the work that earns the mention — content, original data, coverage on the sources the models cite. Divide that spend by the mentions it produced and you get a cost per mention for your brand, which is the closest thing to a price that exists.
Why can't anyone tell me what one mention is worth?
Three reasons. AI answers are generated per request and never published, so there is no impression log to price against. There is no marketplace, so no clearing price forms. And attribution breaks at the source: a buyer who reads about you in ChatGPT and then types your name into Google arrives as branded or direct traffic, and the mention gets no credit for the visit it caused.
How do I estimate the value myself?
Use V × M × C × R × L for the monthly figure and C × R × L for a single mention. Prompt volume and mention rate size the opportunity; consideration lift, close rate and customer value price it. Four of the five terms come from data you already hold. Only consideration lift has to be approximated, and your referral or word-of-mouth conversion rate is the closest proxy.
Is an AI mention worth more or less than a search click?
It is a different unit, which is why comparing them directly misleads. A click buys a session you control; a mention buys a position in a recommendation you do not control. The mention arrives with third-party endorsement a click never has, and it converts more like a referral than an ad. But it may produce no session at all, which means a click-based measurement model will systematically value it at zero.
Does a mention without a link have any value?
Yes, and it is the most commonly undercounted case. The buyer read your name in an answer they trusted and carried it forward, usually into a branded search later. The value is real and the attribution is absent. This is the main reason teams measuring only on last-click conclude AI drives nothing.
How much does it cost to earn an AI brand mention?
It depends entirely on how well the category's cited sources already cover you. The spend is content, original data and earned coverage, not media. The economics differ from advertising in one important way: once the source exists, the mention repeats at no additional cost every time the question is asked, until the model or the source changes. Advertising stops the moment the budget stops.
How many mentions do I need before it is worth measuring?
Measure before you have any. Mention rate at zero is a finding, not a null result — it tells you the model does not consider the brand part of the category, which is the most expensive position to be in and the one most likely to be invisible to every other tool you run.
Can I increase mention value without increasing mention rate?
Yes, and it is often the faster gain. The same mention rate is worth more if you move up inside the answers, get cited rather than only named, or change how the models describe you. A brand named ninth in every answer and a brand named first in the same answers have identical mention rates and very different pipelines.
Start with the term you can measure
Mention rate, share of voice and perception across eight AI models. The rest of the formula is yours.