The 50 questions your buyers ask ChatGPT, Claude, Gemini and Perplexity, checked every week, read by a person, and fixed: pages, corrections at the source, listings and placements. $499 per brand per month.
Shoppers ask the assistants for a product the way they would ask a friend who had done the research: best slim leather wallet, best crossbody under three hundred dollars, which brands use leather that ages well, is this store legit, where is it made. The answer names brands and products, often with shopping cards, and it is written from roundups, retailer product pages and the brands' own sites. For a store, the questions are whether you are named, whether the facts about you are right, and whether your products are in the cards.
Rightcited checks the 50 product and brand questions in your category every week on all four assistants, has a person read every flagged answer, and does the work: the comparison and category pages, the corrections to the records the assistants quote, and the listings on the roundups the answers are written from.
Patterns from the scans we have run this autumn, with the details changed where a company has not agreed to be named.
Rival brands write the roundups the assistants read. Asked which brands use full-grain leather that develops a patina, Claude read six pages and every one was a leather brand's own "best wallets 2026" post; four of the five brands it recommended had written the page they were read from. In a different category the sources are different sites and the pattern is identical. The roundup economy is where product shortlists come from, and a brand that has not written or been included in one is reading about its rivals.
The brand answer turns on the fact your site does not state. Asked whether a leather goods brand with 32,000 on-site reviews was legit, all three assistants said yes and then went looking for where the goods were made. ChatGPT called it "the murky part"; Claude ran five searches and gave up; Gemini read seven of the brand's own pages, found no statement, and wrote "manufactured primarily in China" as fact. A missing fact does not stay missing in an AI answer. It gets filled in.
Records you do not manage write your reputation. The same brand's answer carried a BBB F rating earned by eight unanswered complaints and a 1.6 Trustpilot score under an old domain with 48 reviews, against 32,000 on-site reviews the assistants discounted as the brand's own. The one-year warranty on its warranty page was the pivot of every alternatives answer, read against rivals' lifetime guarantees.
Shopping cards come from catalog feeds. ChatGPT's product carousels in those answers carried rivals' Shopify catalog cards and the retailer listings of brands sold through Nordstrom; the brand's own Shopify store never appeared. OpenAI says product results "are selected independently by ChatGPT and are not ads" and that Shopify merchants are connected through Shopify Catalog; whether a store's catalog is syndicated is a setting, and it is the first thing we check.
Best X for Y and X vs rival pages with the table the roundups lack: leather grade, materials, warranty, price, country of make, shipping, each with its source. One page per question the weekly check shows you absent from.
Where the goods are made, the materials, the tannery or factory, the warranty terms and the return policy, stated plainly on one page so the assistants quote you instead of guessing.
Unanswered BBB complaints answered, review profiles claimed and routed, retailer product pages corrected where they carry a wrong price or spec. We trace each line in the brand answer to the page it came from and fix it there.
Up to 20 inclusion requests a month to the roundups, buying guides and comparison sites the answers read, followed up and logged; up to one secured placement a month where a site will not list you.
Whether your catalog is in the feeds the assistants build product cards from, and what a card needs (price, availability, who makes it) to be chosen; set once, then watched weekly.
Core is $499 per brand per month, month to month, cancel any time, with the work included. The free scan comes first: six to eight of your buyers' questions on two or three assistants, graded by hand, reviewed with you on a 15-minute call. If you start Core, the full 50-question set runs in week one.
Week one: a three-minute intake, then the truth file and the question set within two business days for you to approve. Weeks two to four: the first pages arrive in the order the answers set, the first corrections and listing requests are filed and logged. Day 30: the dashboard shows the first re-check against the baseline, every count opening the answers behind it, with each delivery dated in the work log. Nothing is marked fixed until it holds on two consecutive checks.
$499 per brand per month, month to month: the truth file, 50 questions checked weekly on four assistants, a person grading every flag, up to ten pages, up to 20 listing and inclusion requests and up to one secured placement a month, the dashboard and the work log. Pricing.
The same program at the volume your goals need: more pages, requests and placements a month, community corrections where a site's rules allow, a dedicated channel and a monthly proof call. Scoped and priced after the scan. Book the call.
The AI Page Check runs one page through all four assistants for $2.99, and the free 360-degree scan runs six to eight of your buyers' questions on a 15-minute call. Both are real samples of the weekly check.
Four engagements, four categories, the before and after counts
The ones shoppers ask the assistants in your category: best X, best X under a price, X vs Y, which brands use Z, is this store legit, where is it made, what is the warranty. Fifty, chosen with you; the free scan runs six to eight of them so you can see the method on your own products first.
No. Pages and copy are delivered ready to publish; corrections to third-party records and listings are ours to request and log. The one thing we will ask you to check in your own admin is whether the catalog feed the assistants build product cards from is switched on.
Because the assistants weight independent records over a brand's own, and say so in the answer. A profile with 48 reviews under an old domain can outweigh 32,000 on your site. Claiming and routing those records is part of the first month.
Core at $499 a month covers one brand with up to ten pages, 20 listing requests and a secured placement a month, which fits most single-brand stores. Multi-brand or multi-market accounts are Grow, quoted after the scan.
Or start smaller: the AI Page Check runs one page through all four assistants for $2.99.