
Business case
When AI names your rival: what it costs you
Why a rival named in your place in an AI answer is a lost shortlist you never see, how to count those questions, and how to win them back.
How to put a value on being named in AI answers, using your own pipeline numbers: what to count, the simple sum, and how to show it moving.

Every budget line needs a reason, and AI visibility is a new line. The good news is that you do not need industry averages or a consultant's model to justify it. You can build the case from numbers your team already has, and add one measure that shows the work moving before revenue does.
This post gives you that model: what to count, how the sum works, and how to report it with confidence.
When a buyer asks ChatGPT, Perplexity or Google AI Mode which tools to consider, the answer usually names a handful of brands. Those names become the shortlist. A buyer who never sees your name in the answer may never visit your site, so the loss never shows up in your analytics or your CRM.
AI visibility work pays for two things:
The return is the extra pipeline that comes from both. To measure it, you need one leading number and a few lagging ones.
Revenue moves slowly. A page you publish this month may not show up in closed deals for a quarter. So the model needs a number that moves first.
That number is Mention rate: of the answers read, the share that named you. Ask the same set of buyer questions on the same engines every week, and Mention rate tells you whether you are on more shortlists than last week. Our post Mention rate in AI answers explains how it is counted.
Mention rate is the measure you can act on today. The revenue numbers below tell you what each point of it is worth.
You need three numbers from your own business. Most teams can pull them in an afternoon.
AI engines that link to sources send visitors with a referrer you can see. In your analytics, group the visits whose referrer is chatgpt.com, perplexity.ai, gemini.google.com or claude.ai into one channel called AI assistants. Track it monthly.
These visits are the part you can see. Many buyers read the answer, remember the name and come back later through search or by typing your address, so this channel is a floor, not the whole effect.
Add "An AI assistant (ChatGPT, Perplexity, Gemini or similar)" as an option to the "How did you hear about us?" field on your sign-up and demo forms. Leave it as a choice the buyer picks, not a box you tick for them.
This catches the buyers the referrer misses: the ones who asked an AI assistant, then searched your name. B2B teams have used self-reported attribution for years to catch podcast and community influence that click tracking cannot see. AI answers behave the same way.
Use the figure your finance team already trusts: average first-year contract value, or lifetime value if you report on that. Pair it with your win rate from demo or trial to paid.
With those numbers, the model is short.
AI-sourced customers = AI-sourced sign-ups or demos (from the referrer channel and the form field, without counting anyone twice) × your win rate.
AI-sourced revenue = AI-sourced customers × what a customer is worth.
Return = AI-sourced revenue in a period, minus the cost of the work in that period (the tool, plus the hours spent writing pages and getting listed), divided by that cost.
Here is the arithmetic with made-up inputs, only to show how it runs. Say your forms and referrers show 12 AI-sourced demos in a quarter, you win 1 of every 4 demos, and a customer is worth $9,000 in the first year. That is 3 customers and $27,000. If the quarter's work cost $6,000 in time and tools, the return is $21,000 on $6,000. Put your own numbers in their place.
The sum above tells you what the channel is worth today. Mention rate tells you whether it is growing, and why.
Track both on the same calendar. When Mention rate rises for real on the questions a new page targets, and the AI assistants channel and the form answers rise in the following weeks, you have a story a finance lead can follow: we published, engines started naming us, and buyers started arriving from them.
| Week | Mention rate | Likely range (%) |
|---|---|---|
| Last week | 27.8% | 25.1 to 30.7 |
| This week | 34.0% | 31.1 to 37.0 |
Example, using an invented brand
The likely range around each week matters here. It is the span the true figure almost certainly sits in, and it lets you say "this rise is real" rather than "this looks up". A model built on a number that wobbles is hard to defend. A model built on a number with its range shown is easy to defend. How we measure sets out the method.
A business case for AI visibility fits on one page. Keep it to five lines:
The third line is usually the one that wins the budget. A list of real buyer questions where a named rival appears and you do not is concrete in a way that a market forecast never is. Our post When AI names your rival shows how to put that list together.
The leading number moves first. A new page can be picked up by the engines that search the web within weeks, and Mention rate on its questions shows that. Pipeline follows on your normal sales cycle, so a team with a 60-day cycle should expect revenue to trail Mention rate by about that much.
No. Treat them as the floor. Buyers who read an answer and come back later by search are real, and the form field is how you catch them.
Use the median first-year value rather than the average, so one large deal does not carry the whole case. Report the large deal separately if it came from an AI assistant.
The cost side is small: a plan and a few hours a week of writing. The model works the same at any size, and the questions you lose to a rival are the same questions a buyer is asking whether you track them or not.

Business case
Why a rival named in your place in an AI answer is a lost shortlist you never see, how to count those questions, and how to win them back.

Getting started
What generative engine optimization (GEO) means, how it relates to AEO and SEO, and the five steps that get a B2B brand named in AI answers, in plain words.

Inside BotSay
BotSay's list of pages the engines read when they named a rival: how it is built, what each label asks you to do, and how to work through it.
Your first results arrive right after setup
Monthly plans from $59. Cancel at any time.