The ROI of AI visibility: a model you can defend

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.

On this page
  1. What does AI visibility pay for?
  2. The leading number: Mention rate
  3. The lagging numbers: what to count in your own pipeline
  4. The sum
  5. How Mention rate ties the two together
  6. What to put in the business case
  7. Common questions about AI visibility ROI

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.

What does AI visibility pay for?

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.

The leading number: Mention rate

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.

The lagging numbers: what to count in your own pipeline

You need three numbers from your own business. Most teams can pull them in an afternoon.

1. Visits from AI assistants

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.

2. Self-reported source on sign-up or demo forms

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.

3. What a customer is worth

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.

The sum

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.

How Mention rate ties the two together

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.

Mention rate, two readings
The figures as a table
Mention rate, two readings
WeekMention rateLikely range (%)
Last week27.8%25.1 to 30.7
This week34.0%31.1 to 37.0
Up 6.2 points. A real change.

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.

What to put in the business case

A business case for AI visibility fits on one page. Keep it to five lines:

  1. Where we stand. Our Mention rate across six engines, and our place against our rivals.
  2. What the channel is worth now. AI-sourced sign-ups or demos last quarter, and the revenue they became.
  3. The gap. The questions where a rival is named and we are not, and how many there are.
  4. The plan. The pages we will write and the places we will get listed, in order.
  5. How we will know. Mention rate every week, with its range, and the AI-sourced pipeline every month.

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.

Common questions about AI visibility ROI

How long before the return shows?

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.

Should I count visits from AI assistants as the whole return?

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.

What if our deal sizes vary a lot?

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.

Is it worth it for a small team?

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.

What you gain with BotSay

  • A Mention rate on six engines every week, with its likely range, so the leading number in your case is one you can defend.
  • Your place against every rival you track, on the five measures, for the first line of the business case.
  • The questions where a rival is named ahead of you, grouped by topic, for the gap line.
  • Pages to write, ranked by how many questions each could win, for the plan line.
  • One fixed monthly price with everything included and no add-ons. Plans differ only in the number of questions, from $59 a month for 60 questions.

Keep reading

Your first results arrive right after setup

Find the page to write next.

Monthly plans from $59. Cancel at any time.

Get started