For Shopify stores and any online brand
ChatGPT, Perplexity, Gemini, and Claude recommend products every day. PositionBird tracks whether yours come up, shows you which competitor gets named instead, and writes the content that changes the answer.
> ask: "best soy candles for gifts"
AI Visibility Score: cited by 2 of 4 engines, up 9 points since last scan.
Sample data. Install to run a live scan on your store.
Score, benchmark, fixes and alerts are not four features. They are four turns of the same loop, and each scan starts it again.
Every scan runs your buyer prompts through ChatGPT, Perplexity, Gemini and Claude, plus Google's classic results, several times each. You get one 0 to 100 score and the per-engine, per-prompt record behind it.
For each prompt you lost, PositionBird shows the competitor the engine recommended and the sources it trusted to get there. That turns a low score into a short list of pages to win.
One click drafts the answer-first page, FAQ copy and structured data for the prompt you are losing, written to rank in Google, win the answer box, and earn an AI citation. On Shopify it publishes straight to your store.
The next scan checks whether the fix landed and emails you when a competitor takes a prompt you were winning, with the likely cause. Then the loop starts again from step one.
One click from the App Store. PositionBird asks for the read products permission and nothing else, so it can build prompts from your catalog. It never sees orders, customers, or checkout.
PositionBird suggests the questions shoppers ask AI about your category. Keep the ones that matter, add your own. Typical prompts look like these:
Your 0 to 100 score with a tier, a row per engine and per prompt showing mentioned, cited, or absent, the competitor named instead on every prompt you lost, the sources each engine trusted, and a generated fix for the highest-value gap. Then it re-runs on your plan's schedule.
Every prompt is run multiple times per scan and the results averaged, so a single unlucky AI response doesn’t move your score. The full method is public. Read the methodology
gpt-4.1-miniclaude-sonnet-5gemini-3.6-flashsonarclassic results onlyEvery stored answer carries the model identifier, a timestamp, and a hash. See the published citation study
PositionBird measures all three. Google AI Overviews and AI Mode are a separate surface and are not tracked. See how SEO, AEO & GEO differ
PositionBird measures whether ChatGPT, Perplexity, Gemini, Claude and Google Search results recommend and cite your brand when shoppers ask buyer-intent questions. It rolls the results into an AI Visibility Score from 0 to 100, tracks competitor Share of Voice per prompt, and keeps a citation ledger of the sources each engine trusts.
ChatGPT, Perplexity, Gemini and Claude, each queried through its official API, plus Google's classic results (organic top ten, featured snippet and People Also Ask) for the SEO and AEO layers. Google AI Overviews and AI Mode are not tracked; they are a different surface from classic results and PositionBird does not claim coverage of them. The Free plan checks ChatGPT; Starter covers Perplexity, ChatGPT, Google Search (organic, snippets, People Also Ask); Growth and Scale cover all five.
No. PositionBird is available as a Shopify App Store app and, since September 2026, to any website through web signup with card billing. Web accounts crawl your public site to suggest prompts instead of reading Shopify products.
Yes. The Free plan costs $0, needs no credit card and has no trial clock: 5 tracked prompts checked on ChatGPT with a monthly scan. Paid plans start at $19/month with a 7-day trial.
Three fields: your brand name, your website, and one product category. Nothing else. It takes about a minute, needs no account and no access to your store, and asks five AI engines real buying questions about that category. You see how many engines recommended your store and who they named instead. Enter an email if you want the per-prompt breakdown; it is sent to you and never shared.
Not identically. PositionBird queries each engine's official API, so its answers are a proxy for the consumer chat interfaces, which can vary by session, location and memory. Every prompt is run multiple times per scan and averaged so one unusual response does not move your score.
SEO tools track where your links rank on a results page. PositionBird tracks whether an AI answer names and cites your brand at all, which competitor it names instead, and which sources it trusted, then generates the answer-first pages, FAQ copy and structured data that change the answer.
Step-by-step guides for the day a customer quotes a wrong AI answer at you.