Adobe Expects AI Traffic to Double This Holiday. Four Checks Before November.

The forecast, and what it does not say

Adobe's new holiday forecast, published in late September, expects US online sales to grow 6.7% this season. The line that matters for product teams sits further down: Adobe expects traffic to retail sites from AI chat and browser services to more than double, up 130% on last year, with the biggest jump on Thanksgiving.

Three things keep this honest. First, it is a forecast, built on Adobe's view of more than 1 trillion visits to US retail sites, not an observed result. Second, Adobe counts AI traffic as shoppers clicking a link, and 130% growth describes the rate of change, not the share. For most stores, AI-referred sessions are still a small slice of traffic, and a small number that doubles is still a small number.

Third, the survey findings Adobe pairs with the forecast, that 77% of shoppers who used AI felt more confident in their purchase and 69% said they were less likely to return the item, are self-reported from 5,000 consumers in July. They describe how shoppers feel, not what your returns data will show.

So the question for a Head of Ecommerce is not "Will AI drive our holiday?" It is "If AI assistants send two or three times as many shoppers as last year, are our products in a state to be recommended, and will we be able to tell?" That is a product-information question with a four-week deadline.

What AI assistants read when they recommend

When a shopper asks ChatGPT, Google AI Mode or a similar assistant for "a waterproof hiking boot under $200 for wide feet," the answer is assembled from product information the assistant can retrieve: your product pages, your Google Merchant Center feed, structured data, and whatever third parties say about the product. A product with a complete title, a clear description, accurate attributes and in-stock pricing can be matched to that question. A product whose page never mentions width, or whose feed has a stale price, is easy to leave out of the comparison.

That is why the holiday work is mostly unglamorous. It is not a new channel to launch. It is making sure the products you need to sell in December can be found, understood and trusted by the systems that increasingly sit between the shopper and your site.

Four weeks, four checks

Here is what we check first when a brand asks us to get ready for the season. Each one is scoped so a lean team, or a Paz expert working with that team, can finish it before November 1.

1. Pick the products that carry the season

Start from last year's holiday orders and this year's planned promotions, and list the products and variants that have to perform. For most catalogs this is a few hundred items, not thousands. Everything below is applied to that list first. Fixing the long tail in October is how teams run out of time.

2. Make the feed eligible and current

Open Merchant Center and review the diagnostics for the holiday list. Disapprovals, missing GTINs, price and availability mismatches between feed and page, and products that are limited by missing attributes all take a product out of consideration before any assistant gets to judge it. Fix these at the source, in the platform or PIM that generates the feed, so they stay fixed through the December resubmissions. Where Google has guessed a product category for you, confirm or correct it. Assistants that draw on Google's shopping data inherit those guesses.

3. Close the gaps in product content

For each product on the list, read the page the way a shopper's question would. Does it state the facts a buyer would ask before choosing: size and fit, materials, compatibility, what is in the box, shipping cutoffs for guaranteed delivery? Are the title and description written in the words shoppers use, or in internal SKU language? Are the answers to the common questions on the page at all, or only in a PDF or a support article?

Write the missing facts from your own sources: product specifications, supplier data, support tickets and reviews. If two sources disagree, resolve that with the product owner before publishing anything.

4. Decide how you will know it worked

Before the traffic arrives, set up the measurement. Tag AI referrers as their own channel in GA4 or Shopify, so sessions from ChatGPT, Perplexity, Copilot and Gemini are not buried in Direct or Referral. We walked through the setup in how to measure AI referral traffic in GA4 and Shopify.

Record the baseline now: AI-referred sessions, the landing pages they hit, and conversion rate on those pages for the last 30 days. Then, for a short list of the questions your shoppers actually ask, check whether your holiday products appear in the answers today, and check again after the fixes ship. Appearances in answers, AI-referred sessions and orders are three different numbers. Report them separately and do not let a growth percentage on one stand in for the others.

What to leave until January

A platform migration, a new PIM, a full catalog rewrite and a new structured-data strategy all belong after the season. So does chasing every assistant at once.

If the holiday list is clean in the feed, complete on the page and measurable in analytics, you have done the work that a forecast like Adobe's rewards. If the AI traffic growth arrives, your products are in a position to be recommended. If it arrives slower than forecast, you have still fixed the product pages that every other channel sends shoppers to.

This is the work Paz does for lean ecommerce teams. A Paz expert reviews your holiday products across feeds, product content and AI answers, proposes the fixes with the evidence behind each one, you approve, and the platform publishes the changes and tracks what happens next. If you want a second pair of eyes on your list before November, talk to us.