Back to blog

Enrich Your Catalog Before Black Friday, Not During It

Nov 1, 2026

A dark shop interior just before opening, blinds half raised, with a crowd of small glowing figures waiting outside the glass while the shelves inside are still only half stocked.

There is a version of Black Friday prep every store owner knows: pick the offers, stock the bestsellers, test the checkout, schedule the emails. Shopify's own checklist for this year's sale, published in late October, covers everything from inventory to abandoned-cart emails. Product data gets one line in it: a clear call to action, a product description and payment options on every product page.

That line used to be enough. This year, a growing share of the shoppers who land on your store will have been sent by an AI assistant that read your product pages before they did, and decided which of your products to mention based on what it found there. If it found a title, a price and two sentences, that is all it had to work with.

What the forecasts say about AI shoppers this season

Dozens of thin streams of light pouring down from a cloudy indigo sky and landing on one lit shopfront in a row of darker ones.
The assistants were already sending shoppers last season. This year the forecasts say more of them.

Start with last year. Adobe's recap of the 2025 holiday season puts traffic to US retail sites from generative AI tools up 693% on the year before, and up 670% on Cyber Monday alone. Adobe is careful to add that the base of users is still modest, and that is fair. A small number growing fast is still a small number.

Then look at this year. Adobe's forecast for the 2026 season, covering 1 November to 31 December, expects AI traffic to US retail sites to rise another 130%, with the biggest jumps on Thanksgiving (up 159%), Black Friday (up 95%) and Cyber Monday (up 88%). Adobe counts a visit as AI traffic when a shopper clicks a link inside an AI tool, so those are people arriving at a store, not bots crawling it.

Salesforce's holiday outlook is broader and bolder: it predicts that 20% of all 2026 holiday ecommerce traffic will originate from AI chat agents pulling product data back into conversations with shoppers. Its definition includes the agents themselves, not only the people they send, so do not set it beside Adobe's number. The direction is the same.

And the stakes are not small. Shopify's own figures for the 2025 Black Friday Cyber Monday weekend are a record 14.6 billion dollars in sales from more than 81 million shoppers, peaking at 12:01 p.m. EST on the Friday at 5.1 million dollars a minute. Shopify called it the first holiday season of the agentic era. The assistants were already in the room last year. This year there are more of them, and they have had a year to get better at reading product pages.

Why the week before is the wrong time to fix product data

A long workbench in two halves, the left calm and lit with product cards laid out in neat rows, the right dark and heaped with boxes, tape and tangled cables.
The same work, done in November or done in the week it cannot be done.

Three practical reasons, none of them about AI in particular:

  • Assistants work from what they have already read. Fix a description on the Tuesday before the sale and some assistants will see it that day; others will quote the old one well into December. The earlier the fix lands, the more of the season it covers.
  • Your team is busy with other things. Sale week is ad budgets, stock counts and a support inbox that doubles overnight. Nobody is rewriting 400 descriptions on 24 November, and nobody should be.
  • Change freezes exist for a reason. Plenty of stores stop touching the theme and the catalog in the final days so nothing breaks at the busiest minute of the year. A mass edit of product content the week before is exactly the kind of change the freeze is there to prevent.

So the window is now: the first half of November, before the sale, with time left over to read what changed.

What to fix first, in the order an assistant reads it

  1. Descriptions that answer the question. Gift questions are situational: “a present for a dad who has just started running”. A description that says who the product is for, what it is made of and when you would use it gives an assistant something to match. A spec sheet does not.
  2. Titles and meta descriptions that are unique. If thirty variants share one title, an assistant cannot tell them apart, and neither can a shopper comparing tabs.
  3. Alt text on every image. Assistants and image search both read it, it is required for accessibility anyway, and it is the field most catalogs leave blank.
  4. Structured data that matches the page. Price, availability and brand in markup that agrees with what the page says. During a sale, availability changes by the hour; markup that says in stock under a page that says sold out is the kind of mismatch that makes an assistant stop recommending a store.
  5. Gift collections with a real description. A page that says in plain words who these products are for is the page an assistant can point a broad question at.

Take your top fifty products by last year's sale revenue and do those by hand this week if you have to. Then look at the other 2,950.

Doing it at catalog size, with time to check the work

Hundreds of translucent product cards gliding along a conveyor toward one glowing gate, each card pausing under a small lamp before it passes through.
Drafted for the whole catalog at once, approved one decision at a time.

This is where CortexCommerce fits. It reads your existing product data and images and drafts the first pass for the whole catalog at once: descriptions in your brand voice, unique SEO titles and meta descriptions, consistent tags, and alt text written from what the AI actually sees in each photo. Every draft waits in a review inbox with the before and after side by side and a confidence score, and nothing reaches your live store until you approve it. You can scope a run to one vendor, one collection or one product type, so you can start with the gift-guide products and widen from there.

The part that matters in November specifically: price, SKU, barcode and inventory are locked. The AI cannot write to them, so an enrichment run cannot break the one thing that must not break during a sale. And because the approved copy is written straight back into Shopify or WooCommerce, there is no export, no spreadsheet and no re-import on the Wednesday before Thanksgiving.

Doing this by hand for 3,000 products is weeks of writing. Doing it with a draft-and-review loop is a few afternoons of reading, which is something a busy team can actually finish before the 27th.

We can't promise you a place in anyone's recommendations; nobody honestly can. What we can say is that the shoppers an assistant does send will land on pages that explain the product properly. To see where your store stands before you start, run a free scan at cortexcommerce.ai/scan and see how AI assistants read your products today.

Sources

  1. Adobe Digital Insights: 2026 holiday season online shopping forecast (28 September 2026)
  2. Adobe: Holiday shopping season drove a record $257.8 billion online, with consumers embracing generative AI tools
  3. Shopify News: Shopify merchants generate record-breaking $14.6 billion in Black Friday Cyber Monday sales
  4. Salesforce: Peak holiday 2026 predictions for retail
  5. Shopify blog: Black Friday Cyber Monday checklist for 2026