AI VISIBILITY, MEASURED
ChatGPT is deciding what your customers buy. Are you in the answer?
Shoppers have started asking AI assistants what to buy instead of scrolling Google. We ask ChatGPT, Claude, Perplexity and Gemini the buying questions in your niche — then show you, in one score, whether they recommend you, your competitor, or nobody at all.
01 THE SHIFT
Google gave you ten blue links. AI gives one answer.
When a buyer asks an assistant “what’s the best candle for people with pets?”, it doesn’t list ten stores — it names two or three. If your catalog, feeds and reputation aren’t legible to those models, you don’t rank lower. You simply don’t exist in the conversation.
02 THE SCORE
One number. Six things it actually measures.
Your AI Visibility Score is built from real answers, not guesses: we repeatedly ask the assistants the questions your buyers ask, and weight the result by how purchasing works in your category. Pick a niche — the weights move.
Supplements buyers ask about ingredients and studies; fashion buyers ask about fit and use-case. The score follows the buyer, not a fixed checklist.
03 HOW IT WORKS
Scan. Fix. Watch the number move.
We ask the AIs
Every week we run the buying questions in your niche through ChatGPT, Claude, Perplexity and Gemini — multiple samples each, so one weird answer can’t tank your score.
You get the fix list
A one-click llms.txt, the exact product attributes to add, and a source hit-list — the specific threads, listicles and review pages citing your competitors instead of you.
Monitor & defend
Weekly re-scans chart your share of answer over time. Won a citation? You’ll know. Lost one to a rival? Alert before it becomes a habit.
04 YOUR FIRST REPORT
What lands in your inbox.
05 METHOD
Why one score and not a folder of screenshots.
Repeated sampling
AI answers are noisy. Every prompt runs multiple times per model, every week — your score is a trend line, never a single lucky or unlucky answer.
Real buyer phrasing
Prompts come from how people actually ask assistants for product advice in your niche — use-cases, constraints, and “alternative to X” — not your brand keywords.
Evidence, then fixes
Every point off your score links to the exact answer, source or gap that caused it. No black box, no “trust us, optimize harder.”
Public pages only
The scan reads what any shopper’s assistant can read. No store access, no install, no code on your theme until you choose the monitoring app.
06 FAQ
Asked by every store owner so far.
Yes — the first 100 stores on the waitlist get the full visibility report free at launch: score, share of answer, and the crawl/llms.txt audit. We’re doing it because the first hundred reports double as our calibration set. After that, the scan stays free but the report is trimmed.
Programmatically, the same way a shopper would: buying-intent questions in your category, run through ChatGPT, Claude, Perplexity and Gemini every week. We log whether you’re named, cited, recommended or absent — across repeated samples, so the score is stable.
No. The free scan reads only public pages — your storefront, feeds and the public web. When the monitoring app launches on the Shopify App Store, it will request read-only catalog access, and you can say no and just keep the scans.
You can — once. What you can’t easily do is ask 25–75 questions, across four assistants, multiple times a week, for months, and track which answers change and why. That repetition is the product: one answer is an anecdote, a trend line is a decision.
In batches, in waitlist order. Joining with your store URL puts you in the queue; the moment your batch runs you get the report, and the 1,000-store benchmark dataset lands with batch one.
07 WAITLIST
Find out if the AIs know you exist.
Drop your store URL and email. First 100 stores get the full report free — and the 1,000-store study dataset when it ships.