A holiday 2026 AI shopping playbook for CPG brands

Every holiday season has a new thing that marketing teams are told to prepare for. This year it's AI shopping, and for once the preparation is mostly unglamorous work that also helps the rest of the season: product data, promotions and stock.

This playbook is for CPG brands, the ones selling food, drink, personal care, household and pet products mostly through retailers. We start with the numbers, then lay out what to do and when.

The numbers that matter for CPG

Adobe published its 2026 holiday forecast on September 28, based on more than 1 trillion visits to US retail sites. It covers November 1 to December 31.

Period	Forecast spend	Change vs. 2025
Full season (Nov 1 to Dec 31)	$275.1B	+6.7%
Cyber Week (Thanksgiving to Cyber Monday)	$47.5B	+7.4%
Cyber Monday (Nov 30)	$15.1B	+6.2%
Black Friday	$12.9B	+9.2%
Thanksgiving	$6.9B	+8.5%
Online grocery, full season	$26.1B	+10.3%
Cosmetics, full season	$9.2B	+9.5%

Two things stand out for our readers. Grocery is the fastest-growing category Adobe broke out, ahead of electronics (+5.9%) and apparel (+4.7%). And Adobe says many shoppers will use discounts to stock up on essentials rather than gifts.

Show the data

Category	Expected increase
Clothing basics	+210%
Personal hygiene	+150%
Baby products	+113%
Pet products	+93%
Household cleaning	+49%

Source: Adobe holiday forecast, September 28, 2026

A 150% jump in personal hygiene sales during Cyber Week is a CPG event, whether or not your brand thinks of itself as a holiday brand. Discounts are expected to reach up to 30% off listed price in Cyber Week. And mobile will account for 57.4% of season spending, rising to a record 63% on Thanksgiving.

Where AI fits

Adobe expects AI traffic to US retail sites, measured as shoppers clicking a link from an AI service, to rise 130% year over year this season.

Show the data

Day	Change vs. 2025
Full season	+130%
Thanksgiving	+159%
Black Friday	+95%
Cyber Monday	+88%

Source: Adobe holiday forecast, September 28, 2026

Keep the scale in proportion. AI is still a small share of retail traffic, a point we made in today's think piece. But Adobe's July survey of 5,000 consumers found 77% of people who had used AI to shop felt more confident in their purchase and 69% said they were less likely to return it. Those are shoppers deciding, not browsing.

Three new surfaces will carry much of that traffic this year.

Google's Universal Cart. Launched in the US on October 15 across Search and the Gemini app, it lets shoppers save products from different retailers in one cart. Built on Google's [Shopping Graph](term:Shopping Graph) of more than 60 billion listings, it sends alerts on price drops and restocks and shows relevant promo codes. Checkout runs through Google Pay with merchants including Target, Ulta Beauty, Walmart and Sephora, or passes the cart to the merchant's site.

Promotions in AI Mode. As PPC News Feed reported on October 20, Shopping ads inside AI Mode can show offers like "Buy 4, get $20 off," which shoppers can tap for details. Before, those ads showed only image, price and brand.

Agents that shop. Meta says shopping is the most common thing people use its Muse agent for: about 25% of users, with an average purchase around $200, out of roughly 10 million sign-ups in the US and Canada, according to Payments Dive. Walmart said on October 21 it will launch Agent Profiles within a month, letting customers set spending limits for agents, and require approval for agent purchases by default.

A price drop alert can only fire if the price drop is in the data. A banner on your homepage doesn't count.

The common thread is that all three read structured data. A cart that alerts on price drops needs the price in a feed. An AI Mode ad shows the promotion only if it's set up as a promotion. An agent buying paper towels at Walmart reads Walmart's listing, not your brand video.

The playbook, week by week

When	Focus	What to do
Now to Nov 7	Product facts	Audit top 20 SKUs across your site, Amazon, Walmart, Target and Instacart. Fix size, count, claims and images that don't match.
Now to Nov 7	Feeds	Confirm GTINs, titles and variants in Merchant Center. Apply for or check your ChatGPT feed.
Nov 8 to Nov 21	Promotions	Load Cyber Week offers as structured promotions in Merchant Center and retailer promo tools, with exact dates.
Nov 8 to Nov 21	Answers	Ask Google AI Mode, ChatGPT and Gemini the questions your shoppers ask. Note wrong facts and missing products.
Nov 22 to Dec 1	Stock and price	Keep availability and price current in every feed, daily. Pull promotions the minute they end.
Dec 2 to Dec 15	Review	Re-run the answer check. Fix what was wrong before the last shipping weekend.

Here's the reasoning behind each block.

Now to November 7: get the facts straight

Every AI surface above builds its picture of your product from many sources, and the most damaging problem is a contradiction. If your site says 12 count and the Walmart listing says 10, an assistant comparing price per unit will either pick one at random or skip you. This is the dullest work on the list and the most valuable. Tip 5 has the method.

In the same window, check your feeds. In Merchant Center, make sure every product has a GTIN and that each size and flavor is its own item with its own title. Google's AI surfaces and the cart both draw on the Shopping Graph, which your Merchant Center feed helps fill, which is why tip 16 matters more this year than last. If you sell direct, OpenAI's merchant page lets brands apply to share a feed for ChatGPT shopping; our guide to what's inside ChatGPT Shopping covers what the feed should look like.

November 8 to 21: put the deal where machines read it

Most CPG promotions live in three places: a banner on the brand site, a retailer's promo calendar, and a social post. Only one of those is reliably machine-readable. Set up Cyber Week offers as structured promotions in Merchant Center and in each retailer's own promotion tools, with real start and end dates. That's what lets an AI Mode ad show "Buy 2, save $5" and lets Universal Cart surface a code.

Use this window to see what the assistants currently say, too. Ask the questions your shoppers ask: "best fragrance-free body wash," "is [your brand] good for sensitive skin," "cheapest place to buy [your product]." Write down what's wrong. A short list of fixed errors before Thanksgiving is worth more than a long report after.

November 22 to December 1: keep the data honest

Stock and price change hourly in Cyber Week, and the new tools act on them. A restock alert for a product that sold out again an hour later is a bad experience with your name on it. Make sure inventory feeds update at least daily to Merchant Center and that retailer listings reflect real availability. When a promotion ends, end it everywhere.

If you're deciding whether your own store should let agents buy, tip 33 walks through it. For most CPG brands this season, the agents will be shopping at retailers, and the retailer's rules apply.

December 2 to 15: check and correct

Run the same questions again. Compare the answers with your November notes. Some fixes take weeks to show up as sources are re-crawled, so this is also the time to note what didn't change and plan it for January. For brands with many SKUs, XENA Intelligence is one of the companies building tools for this kind of before-and-after check; a shared spreadsheet works for a short list.

A word on Amazon

Amazon sits slightly apart from all of this. It blocked Meta's Muse agent from shopping its site in September, and it runs its own assistant, Alexa for Shopping, in Amazon's main search bar. For a CPG brand, that means Amazon holiday prep is its own job: the listing, the A+ content, the reviews and the answers its assistant gives. We covered that in our Alexa for Shopping playbook. The short version is the same as everywhere else. Fill in every attribute, keep the facts consistent with your other listings, and read what the assistant says about you before your customers do.

It also means agents like Muse will send holiday shoppers to everywhere except Amazon. If your Walmart, Target or grocery listings have been neglected because Amazon carries most of your online volume, this is the season to fix them.

What we'd skip this season

Not everything labeled "AI holiday prep" is worth doing in a five-week window:

• Rewriting your whole site for AI. Fix the top products. The long tail can wait until January.
• Buying an AI ranking report. Assistants rarely give the same list twice, so a single position number tells you little. We explain why in our post on measuring share of answers.
• Building your own shopping agent integration. Unless you're a large DTC brand, the retailers and platforms are doing this. Your job is to make sure your data on their shelves is right.

The short version

Adobe's numbers say the season will be bigger, earlier and more essentials-heavy than last year, with AI-referred traffic growing fastest from a small base. Google, Meta and Walmart have each launched or announced something since early September that reads your product data and acts on it. None of them can read a homepage banner or a sales deck. Get the facts, the promotions and the stock data right in the places those systems look, and the rest of your holiday plan works better too. Whether you track the results with SearchPanel, another tool or a spreadsheet matters much less than doing the first round of fixes before mid-November.