What this is
A free, client-side AI traffic-at-risk calculator. It produces an illustrative, directional estimate of the visits, customers, and revenue you could be putting at risk when AI assistants and AI Overviews answer your customers' questions without citing you. The estimate is built entirely from inputs you type in, and is returned as a low/base/high range rather than a single number. Updated 2026-06-29.
The formula in plain English
Queries answered by AI = your monthly relevant searches multiplied by your AI-answer-share estimate. At-risk visits = those AI-answered queries multiplied by a capture-loss band — the share that effectively bypasses you when you are not cited — applied at a low, base, and high scenario (30%, 50%, and 70% by default, all editable). At-risk customers = at-risk visits multiplied by your conversion rate. At-risk revenue per month = at-risk customers multiplied by your average customer value, and the annual figure is that multiplied by twelve. Nothing is hidden: there are no baked-in benchmark numbers.
Frequently asked questions
Is this a real measurement of lost revenue?
No — and we will not pretend otherwise. This is an illustrative, directional estimate built entirely from numbers you type in. AI-citation-to-revenue attribution is still immature: there is no reliable, industry-agreed way to measure exactly how much business you lose when an AI assistant answers a question without citing you. Treat the output as a "what-if" model for prioritisation and internal discussion, not as a figure to put in a board deck as fact.
What is "AI answer share" and how do I set it?
It is your estimate of the share of your relevant queries that are now answered directly by AI assistants and AI Overviews — the moments where you would want to be the cited source. Nobody can hand you a precise, audited number for your niche, so you set it yourself. Start conservative. You can sanity-check it by running a sample of your own customer questions through ChatGPT, Perplexity, Gemini, and Claude and noting how often a direct answer appears instead of a list of links to click.
Why does the result show a range instead of one number?
Because a single number would be false precision. Even if your inputs were exact, the share of AI-answered queries that actually bypass you when you are not cited is genuinely uncertain — some users still click through, others act on the answer alone. We apply a wide low/base/high capture-loss band (30% / 50% / 70% by default, and you can change it) and report all three. The honest answer is a range, so that is what we give you.
What assumptions does the calculator make?
Only the ones you can see and edit: your monthly relevant searches, your AI-answer-share estimate, your visitor-to-customer conversion rate, and your average customer value — multiplied by a visible low/base/high loss band. There are no hidden multipliers and no baked-in benchmark numbers. If an assumption looks wrong for your business, change it; the estimate updates instantly.
How do I actually reduce this risk?
Being cited by AI engines comes down to being reachable, readable, and citable: served as real server-rendered HTML (AI crawlers do not run JavaScript), clearly structured, and authoritative on the questions your customers ask. Our free audit samples live answers from ChatGPT, Perplexity, Gemini, and Claude to measure your actual citation rate and hands you a prioritised fix list. See the methodology page for how that measurement works and the sample report for what you get.
What currency does this use?
Whatever you want — the calculator is currency-agnostic. Enter your average customer value as a plain number in your own currency, and the revenue figures come back in that same unit. We deliberately do not attach a currency symbol so the tool works for any market.