How to audit a product catalog for AI shopping visibility before a seasonal push

A seasonal push changes your catalog faster than any other time of year: new SKUs launch, discontinued lines get pulled, prices move daily, and stock levels swing by the hour. AI shopping assistants read all of that through a feed and a set of pages, and a feed that was accurate in September can be stale, mismatched or missing the exact products you are pushing by the time the traffic arrives. This guide is a pre-season checklist for catching that before launch, using Visibility AI’s Website Audit and Monitor layers.

Why this is worth doing before the traffic, not during it

AI assistants are no longer a side channel for holiday shopping. Adobe’s 2026 holiday forecast projects AI-referred traffic to US retail sites will rise about 130% year over year across the November 1 to December 31 season, with Thanksgiving Day traffic up 159%, Black Friday up 95% and Cyber Monday up 88% (Adobe’s forecast, via Digital Commerce 360). Salesforce’s 2026 holiday predictions put AI chat agents at roughly 20% of total 2026 holiday ecommerce traffic, and report that half of shoppers now use an AI assistant somewhere in their buying journey, with 74% saying they trust the product recommendations an AI chat assistant gives them (Salesforce’s predictions, via Communicate Online).

None of that traffic browses your site the way a person does. Botify reports AI bot activity on retail sites has increased more than five times year over year, and the mechanism behind it is a feed check, not a page visit: an assistant reads a product feed or a structured page, not your homepage (Botify). A catalog audit now is the difference between that traffic finding your seasonal lineup and finding whatever stale snapshot an engine last read.

Step 1: Confirm which feeds and pages AI shopping surfaces are actually reading

Start by listing every surface that can source your product data: your Google Merchant Center feed, any feed submitted directly to a platform like ChatGPT’s Instant Checkout, and the product pages themselves, since engines that do not have a feed integration fall back to crawling the page. Auditing your site for AI crawlers covers how to check access at the page level; for a seasonal push, extend that same access check to your feed submission so you know it is current rather than assuming it updates itself.

Step 2: Check that every seasonal SKU resolves as its own product

A new or rebranded seasonal SKU is only visible if an engine can match a buyer’s question to it specifically, not to a parent collection page or a discontinued predecessor. Pull the list of products going into the push and confirm each one has its own page, its own title that states plainly what it is, and a stable product identifier, the same literal-match problem covered in why an AI shopping assistant picks one of your products over another. A seasonal variant folded into last year’s listing as an afterthought reads to an engine as the old product, not the new one.

Step 3: Audit feed attributes against what each engine currently reads

Feed requirements moved substantially in 2026 and a feed built to last year’s spec is missing fields engines now use. OpenAI’s Instant Checkout, built on the Agentic Commerce Protocol, requires a product feed with a stable id, title, plain-text description, link, image, price with currency and current availability, submitted as CSV or JSON and refreshed on a schedule rather than crawled (OpenAI developer docs). Google added a set of conversational attributes to Merchant Center in 2026, including question-and-answer pairs, product highlights and related-product links, built specifically for the longer, conversational queries AI Mode and Gemini Shopping handle, on top of the standard required fields like title, price, availability and image that still decide basic eligibility (ppc.land). Check your feed against both specs rather than whichever one it was built for originally.

Step 4: Verify price and availability freshness under load

Seasonal pricing changes faster than any other part of the catalog, sometimes hourly during peak days, and a feed that updates nightly will show buyers a price or a stock status that is already wrong. Botify’s guidance for this exact problem is to push updated feeds to AI platforms rather than wait to be re-crawled, since a mismatched price shown by an assistant is a trust failure that can reduce how often that product gets recommended at all (Botify). Before the push starts, confirm your update cadence can actually keep pace with how often the seasonal catalog changes, not how often the rest of the catalog does.

Step 5: Run the comparison test on the SKUs the campaign depends on

ChatGPT, Perplexity and Gemini now answer shopping questions with comparison tables that line products up directly against each other, which raises the cost of a thin or inconsistent product page more than a plain list ever did (covered in our piece on the shift to AI shopping comparison tables). Run your highest-priority seasonal queries against your own catalog and a competitor’s before launch, the same way you would track visibility across multiple product lines on an ongoing basis, and fix whichever product is losing the comparison while there is still time.

Audit the catalog while you can still fix it

Run a free Visibility Check against the SKUs carrying the most weight in your seasonal campaign to see whether they are being read, matched and recommended today. Then open Website Audit to check feed and page readiness against current engine requirements, and set up Monitor tracking so the audit becomes a baseline you can watch through the push rather than a one-time look taken too late to act on.

Frequently asked questions

How far ahead of a seasonal push should this audit run?

Four to six weeks before the traffic starts, so there is time to fix feed and page problems and let engines re-index before the surge. Running it the week of Black Friday only tells you what is already broken with no runway to fix it.

Does this replace ongoing AI visibility tracking?

No. This is a one-time readiness check before a specific spike in demand, not a substitute for tracking presence and share of voice over time. Run it as a checkpoint inside your regular monitoring, not instead of it.

What if only some products are part of the seasonal push?

Audit those first and treat the rest as lower priority. A catalog-wide audit is worth doing eventually, but the SKUs actually driving the campaign are where a feed gap costs you real traffic, so work outward from there.