My customers won't shop with an agent. Here's why that won't last

Will customers use an agent?

Right now, fewer than will in a year, but the direction is clear. In the UK, Adyen's 2026 retail report found the share of shoppers using AI for shopping rose from 12% to nearly 28% in twelve months, and 44% said they'd let AI handle a whole purchase once they'd set a budget and criteria.

The US shows where this goes, because the services have been live longer. In Adobe's survey of over 5,000 US consumers, reported in April 2026, 39% already use AI for online shopping and 85% of them said it improved their experience. AI traffic to US retailers rose 393% year on year in the first quarter of 2026, and converted 42% better than other traffic in March.

So if your customers say they'd never let a robot shop for them, that's a fair description of today. If the US is any guide, it's a poor description of where the UK is heading.

Does it matter if my regulars never use one?

Because the question changed underneath the argument. Nobody has to hand over a card for an assistant to cost you a customer. They only have to ask it where to look.

In a survey of US consumers by L.E.K. Consulting, reported by MarTech in July 2026, 46% of AI users started their purchase research on a standalone AI platform, up from 25% in 2024. And 31% said their decision was largely made before they reached a brand or retailer site.

Your customer may still pay on your site, by hand, with her own card. She just might not have found you there.

What does our own research show?

On 26 October 2026 we asked six AI assistants for the best brands in their category, for 75 well known UK brands across seven sectors.

Fifteen of the 75, one in five, weren't named by any of the six. Another 15 were named by only one, so 30 of the 75, or 40%, were close to invisible. Only 18 were named by all six.

Being established doesn't protect you. Five of the 12 fashion and footwear brands were named by none of the assistants. In sport and outdoor retail, the average brand was named by one of the six. And the assistants disagree: when one recommends a brand, on average 39% of the others leave it out.

These are real businesses with real shops and real revenue, and their customers aren't asking an agent to pay for them either. They may be asking one who to trust, and the answer may not include them. It's one run, in one market, on one day, so treat it as a snapshot rather than a verdict. The full report and method are public.

Haven't we heard this before?

Yes. When online shopping arrived, plenty of successful shops said their customers liked to see things, wanted to talk to someone, and would never type a card number into a computer. They were right about their customers at the time. They were wrong about where it was heading. In June 2026, 29.4% of all UK retail spending was online, according to the ONS figure reported by Parcelhero.

Customers didn't change overnight. The default did. Three things were true then, and we think they're true now:

1. New customers move first.Your regulars may never change. The people who haven't met you yet start somewhere else, and they are next year's regulars.
2. Nobody decides to switch.No one announced that they were shopping online now. They started where it was easier. Assistants are being built into search, phones and browsers: Google says AI Mode passed a billion monthly users, and Apple said in January 2026 that its next foundation models would be based on Google's Gemini. People will use them without choosing to.
3. Late is expensive.The shops that treated a website as a side project spent years catching up with the ones that didn't. The catch up cost more than the early start would have.

What does “you won't have a choice” actually mean?

Not that you must sell through an agent. You can keep your own checkout, your own shop and your own way of doing business, and we think many good brands will.

It means you don't get to choose whether assistants talk about you. They already do, from whatever they can read. A brand that is readable and chooses not to sell through agents loses nothing. A brand that is unreadable and never looked loses the shopper who never found out it existed.

Adobe measured AI referral traffic to US retail sites up 62% year on year in July 2026, converting 60% better than other traffic.

What should a successful retailer do this quarter?

Nothing dramatic. Three steps.

1. Look before you decide.Ask the assistants the questions your shoppers ask. Our free score does it across eight engines and every market you sell in, and shows what they say.
2. Fix the dull thing.Product data. Titles, sizes, colors, prices, stock and delivery times, accurate and current. It's what every assistant reads first, and it helps your own site too.
3. Keep your own checkout and your own customer.Let assistants send people to you. Then give them a reason to come back that doesn't depend on an assistant: an account, email and a community of their own.

What happens next?

Today's customers may well say no, and that's fine. The shift won't come from persuading anyone to trust a robot with their money. It will come the way online shopping did, from a new default that slowly became the first place people looked, and for many US shoppers it already has. The brands that treat it as someone else's problem will find out when the new customers stop arriving. And if a few assistants do end up dominating, regulators have a record of insisting on fair access, which is why being readable to all of them matters more than picking one. We set out how we expect that to play out.