Most of the conversation about AI shopping focuses on ChatGPT. But there's a second assistant quietly sending shoppers to stores every day — and it works differently enough that being visible in one doesn't mean you're visible in the other.
That assistant is Perplexity. If you've been optimizing only for ChatGPT, you may be invisible in Perplexity without realizing it — or, occasionally, the reverse. Here's how Perplexity decides what to recommend, why it's not the same as ChatGPT, and how to check where your Shopify store stands.
What makes Perplexity different
ChatGPT, at its core, generates an answer from what it has learned, then increasingly backs it with live search. Perplexity is built the other way around: it's a search-first answer engine. Almost every response is assembled from live web sources it retrieves in real time, and it shows those citations right in the answer.
That single design difference changes everything about how you get recommended:
- Citations are everything. Perplexity leans heavily on the specific pages it can find and cite for a query. If your store — or an article mentioning your store — isn't in the sources it pulls, you won't appear, no matter how good your products are.
- Freshness matters more. Because it retrieves live, recently published and recently updated content carries more weight than it might in a model relying on older training data.
- The sources it trusts are often third-party. Review sites, comparison articles, "best of" roundups, and reputable editorial pages frequently outrank a brand's own product page as a citation. Being talked about elsewhere matters as much as what's on your own site.
The practical upshot: a store can rank fine in ChatGPT (which may recall it from training) yet be missing from Perplexity (which only surfaces what it can cite right now) — and the fix for each is not identical.
How to check whether Perplexity recommends you
The test mirrors the one you'd run for checking whether your store shows up in ChatGPT, but do it separately, because the answers genuinely differ.
- Open Perplexity in a logged-out or private window.
- Ask the way a shopper would: "What are the best [your category] brands to buy online in the US?" and a couple of variations.
- Read the answer and the citations. Note two things: is your brand named, and which sources is Perplexity citing to build its recommendation?
That second part is a gift Perplexity gives you that ChatGPT often doesn't — it shows its work. The cited sources are a literal to-do list: those are the pages you'd want to be mentioned on to enter the answer.
Why you might be in one but not the other
If you find you're recommended in ChatGPT but not Perplexity, it usually means your brand has some presence in the model's training or memory, but not enough current, citable web coverage for a search-first engine to surface you. The fix is external: get mentioned, reviewed, and compared on pages Perplexity can cite.
If it's the reverse — Perplexity names you but ChatGPT doesn't — you likely have decent recent web coverage, but not the broader, repeated presence that a training-based model accumulates over time. That's a slower build.
Either way, the lesson is the same one that keeps coming up in AI-era discovery: AI recommends what it can verify about you elsewhere, not simply what's best. And "elsewhere" is weighted differently by each assistant.
Don't optimize for one assistant and call it done
The biggest mistake here is treating "AI visibility" as a single number. It isn't. ChatGPT, Perplexity, Gemini, and Claude each pull from different places and update on different schedules. You can be a default answer in one and completely absent from another — and if you only ever check one, you're flying blind on the rest.
This is exactly why measuring visibility per assistant, rather than in aggregate, matters. Vanto checks each of your products across ChatGPT, Claude, Gemini, and Perplexity separately — so you can see, engine by engine, where you're recommended, where you're invisible, and which competitors are taking your place. And for stores already getting AI-driven orders, it attributes real Shopify revenue to the channel that drove each one.
Start with the free test above in Perplexity today. Then check ChatGPT. If the answers differ — and they usually do — you've just learned that "being visible in AI" is really several different jobs, not one.