When a buyer asks ChatGPT or Perplexity which tool to use and your product is one of the names that comes back, that's a high-intent citation. They're not browsing, they're choosing. Put your own subscription price in and you'll know what one is worth to you.
Guessing at your citation count? I'll find the real one. Send me your product and I'll run the queries your buyers actually ask, count where you show up, tell you who's being cited instead, and map what it'd take to displace them. Free.
It's a model, not a measurement. AI citation behaviour is new, unreported by the platforms, and varies by category and query type. It counts only high-intent citations and ignores educational queries, brand recall, competitive displacement, and the compounding effect of being the cited source over time, all of which push the real figure up. Use it to size a decision, not as a forecast.
No black box. Every step's visible, and every assumption above the fold is adjustable. If you disagree with one, move it and watch the number change.
Your monthly subscription price multiplied by average customer lifetime. Lifetime's derived from churn: a 3.5% monthly churn rate, the B2B SaaS median, implies a customer who stays roughly 29 months.
Not every citation becomes a customer. Some people who see your product named in an AI answer click through, some of those start a trial, and some of those convert to paid. Multiply the three and you've got the odds that any single citation eventually produces a customer.
Lifetime value multiplied by those odds gives the expected value of a single citation. At a $49 product with median churn, that lands around $7. It sounds small until you multiply it.
Your monthly citation count times the value of one, times twelve. That's the annual figure. Unlike paid acquisition, the assets earning those citations keep earning them after you've stopped spending.
By the time someone clicks out of an AI answer, the research is over. The model has already defined their problem, compared the options, and told them you're one of the answers. They don't arrive to browse. They arrive to verify. That pre-qualification is the whole premium, and it only applies to brands the model actually cites.
More than half of AI-referred sessions land straight on a product page, against roughly 20% for organic search. They skip the awareness content entirely, which is why the conversion gap shows up.
This evidence is contested and early, and you should know that before quoting any of it. The largest study to date, covering 973 ecommerce sites and over 50,000 ChatGPT transactions against 164 million traditional ones, found the opposite: ChatGPT referrals converted worse than organic search and produced lower revenue per visit. A paired analysis by Amsive found no significant overall difference across 54 sites, with the B2B advantage above holding only in the B2B subset.
Volume is also small. LLM referrals are still around 1% of sessions for most B2B sites. What's not in dispute is the growth rate, with AI referral traffic up several hundred percent year on year, and the mechanism, which is that a cited brand arrives pre-qualified in a way a cold search click never does.
The B2B picture is consistently better than the B2C ecommerce picture, which is the segment most of the sceptical data comes from. That's the honest read: promising, directionally clear, not yet settled.
Most AI visibility tools report one number: how many times your brand got mentioned. That number's close to useless on its own, because a mention inside a general explainer and a mention inside "best tool for X" are worth wildly different amounts.
"Best guest verification software", "Product X vs Product Y", "software for short-term rental hosts", "alternatives to [competitor]". Someone's choosing a tool right now and your name's in the shortlist. That's the citation that converts.
"What is a digital guidebook", "how does guest verification work". Real value, much slower payoff, and nowhere near the same conversion rate. Including it'd inflate the number.
Your brand named in passing inside an answer about something else entirely. Nice for recall. Worth roughly nothing on its own.
That's why the figure here usually comes out lower than tools that count every mention. That's deliberate. A number you can defend in front of a CFO's more useful than a big one you can't.
Models cite what they can verify, attribute and quote. In practice that narrows to a short list, and most content marketing doesn't qualify.
An AI citation is when a large language model names or links to your product inside a generated answer. It happens in ChatGPT, Perplexity, Google AI Overviews, Gemini and Copilot. Unlike a search ranking, the user often never clicks, so it can influence a purchase without ever appearing in your analytics.
A citation on a query where someone is actively selecting a tool, rather than learning about a category. "Best project management software for agencies" is high intent. "What is project management software" isn't. This calculator counts only the former, because the conversion difference between the two is large.
Customer lifetime value multiplied by the probability that a single citation eventually produces a paying customer. That probability is the product of three rates: citation to site visit, visit to trial, and trial to paid. All three are adjustable in the calculator.
Because no AI platform reports citation click-through. There's no equivalent of Search Console for ChatGPT. Churn and trial-to-paid rates come from published B2B SaaS benchmarks. Citation-to-visit and visit-to-trial are informed estimates, set for high-intent queries where the user is evaluating rather than browsing. Anyone offering you a precise figure on this is guessing.
With difficulty, which is the honest answer. No AI platform reports citations the way Search Console reports impressions, so there's no dashboard to check.
Three options. Paid tracking tools like Otterly, Profound, Peec and Scrunch monitor a set of prompts across engines and report where you appear. Manual checking works too: write down the fifteen questions a buyer would genuinely ask when choosing a tool in your category, run each through ChatGPT, Perplexity and Google, and count how often you are named. Tedious, free, and more accurate than a guess.
Or send me your product and I'll do the manual version for you at no cost, which is the offer above.
It varies enormously by category size and existing authority. A handful of well-built assets, meaning one research report with supporting articles and an explainer video, commonly produces citations in the low hundreds per month in a mid-size category. A single strong explainer video can carry a meaningful share of that on its own. A brand new domain in a competitive category'll take considerably longer.
Video punches above its weight. YouTube is consistently among the most-cited domains in AI answers, ahead of most publishers, and video is a small fraction of total content. The format matters though. In testing across a live B2B SaaS channel, plain explainer videos earned citations while research-led videos earned almost none, because explainers answer the question a buyer actually types when they're choosing a tool.
In B2B, the evidence points that way. Ahrefs found ChatGPT drove about 0.5% of their visits but roughly 12.1% of signups. Semrush put AI search visitors at 4.4 times the value of organic. Seer Interactive measured ChatGPT referrals converting at 15.9% against 1.76% for Google organic. Amsive found LLM traffic converting at 2.17% versus 1.16% for organic in the B2B subset of their data.
But it's contested. The largest study, across 973 ecommerce sites, found ChatGPT converting worse than organic. Amsive found no significant difference overall. Most of the sceptical data comes from B2C ecommerce, where the pattern genuinely looks different. Volume is also still around 1% of sessions for most B2B sites, so this is a high-quality trickle rather than a flood.
Most AEO ROI calculators start from your investment and project a return, which requires you to already know what you'll spend. This starts from the unit: what is one citation worth. That makes it useful before you've decided on any budget, and it's deliberately conservative, counting only high-intent citations and ignoring brand recall entirely.
No. The calculator's open, nothing's gated, and no data leaves your browser. There's a form further down if you want a citation plan for your product, but it's entirely optional.
The longer version, if you'd rather give me more context. I'll map the queries you should be showing up for in AI answers, who's getting cited instead of you, and the research and video that would put you there. Specific to your product, at no cost.
Prefer to talk first? Book a free 30-minute call · Connect on LinkedIn