AI answers change every time. What am I paying to fix?
They do change. In SparkToro's study of almost 3,000 runs, ChatGPT and Google's AI gave the same list of brands less than 1% of the time. In our own Thanksgiving series, the same question asked two hours apart kept one brand out of six on Day 1.
What holds steady is how often a brand gets named across many runs. In the same study, some brands showed up in 60% to 90% of answers to the same kind of question, run after run. That share moves when you fix what AI reads: your product pages, product data, reviews and retailer listings. So the audit measures your seat rate, the share of answers that name you, and re-tests it 30 days after your fixes.
Back to the questions
Isn't AI traffic too small to matter yet?
For most stores it's still a small share of visits, and anyone who tells you otherwise is overselling. Two things make it worth checking now. Shoppers who arrive from AI convert well: Adobe found AI visitors to US retail sites converted 90% better than other visitors in September 2026. And shoppers already ask AI about brands by name, so what it says about you reaches people who were about to buy.
The audit tells you whether there's a gap worth fixing before you spend more on it.
Back to the questions
Isn't this just SEO with a new name?
Partly, and that part is good news. Google says its AI Overviews and AI Mode have no extra requirements beyond regular Search, so most fixes help both.
ChatGPT reads a wider set of pages. In Ahrefs' study of 15,000 prompts (August 2025), only about 8% of the pages ChatGPT cited ranked in Google's top 10 for the same question. It leans on lists, reviews, forums and retailer listings, so a brand can rank well on Google and still go unnamed in ChatGPT. The audit shows which pages each engine reads for your buyers' questions.
Back to the questions
Can't I check it myself?
Yes, and you should. Open ChatGPT in a private window, stay logged out, ask what your buyers ask and write down every brand it names. The free check writes three of those questions for you.
The audit is the version that takes days: 30 questions on four engines, each asked three times at different times of day, with the sources each engine used, the brands named in your place and the fixes ranked by impact.
Back to the questions
Can you guarantee AI will recommend us?
No. Nobody controls what AI says, and a guarantee of a ranking or a spot in ChatGPT is a promise nobody can keep.
Here's what I do guarantee: your report arrives on the date I give you before you pay, or it's free. It has at least three fixes your team can act on, or you get your money back. And your fixes get re-tested 30 days later on the same questions, so you can see what moved.
Back to the questions
You're one person. What happens if you get busy?
I run every audit myself, so the person on your results call is the person who tested your brand. To keep that true, I take two audits a week and two Program clients at a time. If I'm full, I'll give you the next start date before you pay.
Every fix is written down in your report, so your team or agency can carry it out with or without me.
Back to the questions
Who else have you done this for?
No ecommerce client yet, and I'd rather say so plainly. The first two brands are picked on November 20. Each one gets re-tested after 30 days, and the before-and-after goes on this site in early January with the brand's permission.
Until then, the proof is public: the Thanksgiving series, with every run dated, and my own scorecard, including where AI got my own business wrong.
Back to the questions
Why not just buy a tracking tool?
A tracker is useful once you know which questions matter and what to change. It runs your prompts on a schedule and gives you a number.
The audit gives you the reasons behind the number: which sources each engine trusts, who gets named in your place and what to fix first. Most brands need the audit first and tracking after.
Back to the questions
It's Q4 and we're slammed. Can this wait?
It can, and the audit is built for that. It needs nothing from your team during the holidays: I only need your website, your products and three competitors.
You get a ranked list to act on in January, when next year's budget gets set. Many fixes are text your team edits in your store admin, with no code release.
Back to the questions
How do you run the questions?
Every question is asked three times, at different times of day, on ChatGPT, Google's AI, Perplexity and Claude. A 30-question audit comes to 360 answers.
The screenshots are my own runs in a logged-out private window, the way a shopper sees it. The repeat runs go through each engine's official developer access, so they show how often a brand comes up across many answers. Developer answers can differ from the app, which is why every screenshot in your report comes from the shopper-facing app. Your report labels which is which, and everything is dated.
Back to the questions
What do I get, and what does it cost?
The AI Visibility Audit. 30 of your buyers' questions on four engines, each asked three times. You get your seat rate on each engine, what AI says about you by name, dated screenshots of the key answers, the brands named in your place, at least three ranked fixes and a 30-minute results call. It's $500 (£400 for UK brands) for orders through November 30, then $995 (£750).
The Visibility Program, after the audit. 90 days: up to five pages rewritten each month, a fix checklist for your team, a re-test every month and one call a month. $3,000 paid upfront, or $1,100 a month for three months. Your audit fee counts toward it. After the 90 days, keeping your seat is $500 a month.
Back to the questions
Sources