AI assistants will send 130% more shoppers to US retail sites this holiday season than last, according to a forecast Adobe published on 28 September 2026. The figure counts clicks from AI chat and browser services, and it gives UK ecommerce teams about eight weeks before Black Friday on 27 November to measure the channel properly.

Adobe’s numbers for November and December

Adobe’s holiday forecast expects US online spending to reach $275.1 billion between 1 November and 31 December, up 6.7% on 2025. The AI figures sit inside that. Adobe defines AI traffic as shoppers clicking a link from an AI service to a retail site, and it expects the steepest rise on Thanksgiving (up 159% year on year), with Black Friday up 95% and Cyber Monday up 88%.

The forecast draws on more than 1 trillion visits to US retail sites and 100 million products across 18 categories, backed by a survey of 5,000 consumers in July. Among people who had already shopped with AI, 77% said it made them more confident in a purchase and 69% said they were less likely to send an item back. Adobe says it first saw shoppers using AI chat and browser services for deals and product searches in the 2024 holiday season.

Vivek Pandya, Adobe’s director of digital insights, said consumers are “getting much more comfortable having dialogues with an agent, getting recommendations and going to the retail site from those conversations”, and that the use now reaches “beyond big-ticket items to everyday essentials”, as the Spokesman-Review reported.

Why a US forecast should change a UK plan

Adobe’s data is American and Thanksgiving means little to UK shoppers, but Black Friday and Cyber Monday are fixed points in the UK retail calendar, and the assistants driving the US numbers are the same ChatGPT, Gemini, Perplexity and Copilot products British shoppers use. A UK brand that can’t see its own AI referrals today will go into the busiest trading weeks of the year without a baseline.

US shoppers are handing product research to these assistants. When a shopper asks an assistant to compare three running jackets under £150, the assistant reads product pages, reviews and retailer listings, then sends a click to whichever page answered best. Pages with thin descriptions, sizing hidden in images or prices that load late give the assistant less to quote. Your product catalogue now does part of the job a comparison site or a paid search ad used to do.

Our view at Manual Focus is that UK retail and DTC brands should treat AI referrals as a named channel in their peak trading report this year, with its own baseline and targets, even while the numbers are small. We’d set the targets on conversion and returns rather than volume, because Adobe’s survey points at a shopper who arrives further along in the decision. Paid placement in these assistants is arriving too, which we covered in what ChatGPT Ads on Shopify means for UK ecommerce brands.

Four jobs to finish before 27 November

  1. Build an AI referral channel in analytics this week. Group traffic from chatgpt.com, perplexity.ai, gemini.google.com and copilot.microsoft.com into one channel, then pull September and October as your pre-peak baseline.
  2. Compare it with paid search and organic. Put conversion rate, average order value and 30-day return rate for AI referrals next to your other channels. The attribution tear-down playbook covers how to avoid double counting when the same shopper also clicks an ad.
  3. Fix the pages assistants would quote. Take your 20 best-selling Black Friday products and make sure sizing, materials, compatibility, price and stock appear in plain text on the page.
  4. Check nothing blocks the assistants. Ask whoever runs your CDN and robots.txt to confirm that bot-protection rules haven’t shut out the AI search crawlers that fetch your product pages.

Gaps in the forecast

Adobe’s release gives growth rates but not the share of all retail visits that AI referrals make up, so a 130% rise could still be a small slice of traffic. The count covers clicks only, which misses every purchase an AI answer shaped without sending a visit. The survey is self-reported and limited to people already shopping with AI. And the forecast covers only the US, so British brands have to build the evidence from their own data.

If you want help setting up the peak trading report, get in touch.

Frequently asked questions

How much will AI shopping traffic grow this holiday season?

Adobe forecast on 28 September 2026 that traffic from AI services to US retail sites will rise 130% year on year over November and December, measured by shoppers clicking a link. It expects the biggest jump on Thanksgiving, up 159%, followed by Black Friday at 95% and Cyber Monday at 88%. The forecast covers the US only.

Do shoppers trust AI assistants for buying decisions?

In an Adobe survey of 5,000 consumers fielded in July 2026, 77% of people who had used AI for online shopping said it made them more confident in their purchase, and 69% said they were less likely to return an item. The survey is self-reported and covers people who already shop with AI, not all shoppers.

Does Adobe's AI traffic forecast apply to UK retailers?

Not directly. Adobe's 2026 holiday forecast uses US data only, drawn from more than 1 trillion visits to US retail sites. UK brands should read it as a signal of where shopper behaviour is heading and check their own analytics for AI referrals, rather than applying the 130% figure to a UK plan or budget.

How do I track ChatGPT and Perplexity traffic to my ecommerce site?

Create a custom channel group or segment in your analytics tool that matches referrals from the main AI assistant domains, such as chatgpt.com, perplexity.ai, gemini.google.com and copilot.microsoft.com. Review the list monthly because new referrer domains appear. Adobe counts AI traffic by clicks, so this captures visits but misses purchases influenced by an AI answer without a click.

Sources

From the podcast

James and Hannah talk through Adobe’s holiday forecast of 28 September 2026, which expects clicks from AI services to US retail sites to rise 130% year on year over November and December, with Thanksgiving up 159%, Black Friday 95% and Cyber Monday 88%. They cover what Adobe counts as AI traffic, how much weight to put on its survey, why a US forecast should still change a UK peak trading plan, and three jobs to hand to AI before Black Friday on 27 November.

Get AI on this

Paste these into whichever AI assistant you use: ChatGPT, Claude, Gemini or Copilot. Swap the bracketed parts for your own details. The first two need file uploads, and the first should be set to run monthly on an assistant that can work on a schedule. The third needs an assistant that can browse the web.

1. Build the AI referral channel and baseline

I run ecommerce marketing for [brand], [one line on what you sell] at
[website], selling mainly in [market].

I want to treat traffic from AI assistants as a named channel in our peak
trading report, with a pre-peak baseline from September and October.

Attached is a referral export from our analytics for [date range], with
referrer domain, sessions, conversions and revenue by week.

1. Pick out every referral from an AI assistant, including chatgpt.com,
   perplexity.ai, gemini.google.com and copilot.microsoft.com, and any
   other referrer you're confident is an AI assistant.
2. Give me a weekly baseline table with the columns: Week | AI referral
   sessions | Conversions | Revenue | Share of all sessions.
3. Write a single matching rule (a regular expression on referrer domain)
   my analytics team can use to set up a custom channel group.
4. List any referrer domains you think might be AI assistants but aren't
   sure about, so I can check them.

Use only the numbers in the export. If a figure isn't there, write "not in
data". Note that this counts visits only; it can't capture purchases an AI
answer shaped without a click.

Then set this up as a routine that runs on the first Monday of each month.
Each time, I'll attach the latest referral export. Update the baseline
table, and compare this month's AI referrer domains with last month's and
tell me which are new.

2. Compare AI referrals with paid search and organic

I run ecommerce marketing for [brand] in [market]. We've set up AI
assistant referrals as their own channel in analytics and want to judge
them on conversion and returns rather than volume.

Attached are:
- a channel export for [date range] with sessions, conversion rate and
  average order value for AI referrals, paid search and organic search
- a returns report showing the 30-day return rate by acquisition channel

Give me one table with the columns: Channel | Sessions | Conversion rate |
Average order value | 30-day return rate.

Then, in no more than five sentences, say how AI referrals compare with the
other two channels on conversion and returns. Flag anywhere the same
shopper may have been counted in more than one channel, for example someone
who arrived from an assistant and later clicked a paid ad. If the data
doesn't let you check that, say so.

3. Check the pages assistants would quote

I run ecommerce for [brand] at [website]. When a shopper asks an AI
assistant to compare products, it reads product pages and sends the click to
whichever page answered best. Thin descriptions, sizing in images and
prices that load late give it less to quote.

Here are our 20 best-selling Black Friday products: [paste product names
and URLs].

For each URL, open the page and check whether these appear in plain text
on the page (not only in an image, a tab that needs a click to load, or a
PDF): sizing, materials, compatibility, price, stock.

Give me a table with the columns: Product | URL | Sizing | Materials |
Compatibility | Price | Stock, marking each "plain text", "image only",
"missing" or "couldn't check".

Then, for every product with a gap, draft the missing copy in no more than
60 words, using only details already on the page. Where the detail isn't on
the page at all, write "needs input from [team]" instead of inventing it.

And one for you

Ask whoever runs your CDN and robots.txt to confirm that bot-protection rules haven’t shut out the AI search crawlers that fetch your product pages. It’s a quick conversation, and it’s worth having before 27 November.