Questions owners ask

When shoppers ask ChatGPT, Gemini or Claude what to buy (“best cast iron skillet”, “X vs Y”, “gift for a coffee nerd under $100”), the assistant answers with a short list of products. AI visibility is whether your brand is on that list, where it appears, what the assistant says about you, and which sources it cites to back the answer up.

ChatGPT, Gemini and Claude, each with web search turned on, because that is how they answer shopping questions. Perplexity is next. Microsoft Copilot and Google AI Mode don’t offer a way to ask them questions programmatically, so we cross-check those with the AI performance data in your Shopify admin and with manual spot checks.

Since March 2026, Shopify sends eligible US merchants’ products to ChatGPT, Copilot, Gemini and Google AI Mode by default. That puts you in the catalog. It doesn’t decide whether an assistant recommends you when a shopper asks for the best option.

In our pilot, seven of sixteen Shopify brands were named in fewer than half of the AI answers in their own category, while the category leaders were named in most of them. Being listed is the starting line. Being recommended is the work.

It overlaps, but the goal is different. A search engine ranks a list of pages. An assistant writes one answer, recommends a few products and leans on a handful of websites. We check that answer directly, improve the product information and pages assistants read, and work on getting you covered by the websites they rely on. We show you the raw questions and answers so you can judge the difference yourself.

No, and be careful with anyone who does. Nobody outside the AI companies controls how an assistant ranks products or what it learned in training. We guarantee the work: the measurements, the fixes, the reports and the outreach we agree on. We report results honestly, including when nothing moved.

We ask each assistant the same questions every month and save the full answers: which products it recommends, in what order, and which websites it relied on. Any single answer can change from one day to the next, so we report the trend over repeated checks rather than one screenshot.

We ask through each assistant’s developer access, which can differ slightly from what a shopper sees in the app, so we also spot-check by hand and compare with the AI data in your Shopify admin.

A short report, reviewed by a person: ten questions your customers might ask, put to ChatGPT, Gemini and Claude; which brands they recommend; where you show up, if you do; and the three gaps we’d fix first. Request it with your store’s address. No call and no obligation.

How often you were recommended this month compared with last, which competitors and websites are winning your customers’ questions, what we changed and why, any mistakes assistants made about your store, and what we’ll do next. Tara writes it up and is available for a call. See the sample report.

For the snapshot, just your store’s address. For paid work: a Shopify collaborator account with limited permissions, the competitors you lose to, the products you care about most, and your sign-off on changes. On the Growth plan, we apply approved changes ourselves.

Catalog and on-site fixes can be live within days, and assistants that browse the web may start using them within weeks, though there’s no fixed timeline. Getting into the review sites and roundups that assistants cite takes longer, usually months, because it depends on editors and reviewers. That’s why paid plans start with a 90-day term.

No. Software does the repetitive part: asking the assistants hundreds of questions and reading every answer. A person does the rest. Tara checks each finding against the saved answers, edits every report before you see it, decides what to fix first, and handles the outreach. You get a person to email and talk to, not a login and a chart.

Brandon is the name of our AI analyst. It asks the assistants your customers’ questions, reads every answer and drafts what it found. It’s software, not a person, and it never sends you anything on its own: Tara reviews and edits everything first.

We don’t write fake reviews, post as fake customers in forums, hide text or instructions on your site aimed at AI systems, or promise placement. Fake reviews are also illegal under FTC rules. More on how we work.

Paid plans have a 90-day initial term, then continue month to month. The baseline audit is a one-time $750 and is credited to your first month if you continue. Our first three clients get 50% off for 90 days in exchange for a case study. See pricing.

The questions we ask AI assistants are about your category and products, which is public information. If you give us access to your store, we use it only to measure and to make the changes you approve. We don’t sell data. Details are in our privacy policy.