How AI Agents are changing the shopping journey

With Artificial Intelligence penetrating almost every aspect of our lives right now, it was just a matter of time when it will start assisting customers' shopping journey.

Let's picture a situation: you want to buy a new swimming suit for the vacations. It needs to meet your specific criteria such as colour, size, cut and sustainability. You ask an AI chatbot, it analyses your input, your previous purchase history, and products that fit your criteria, and gives you links to about 3 products while explaining why those are the perfect fit for you. That is how AI shopping assistants work, and it is already changing what retailers have to get right to be found, trusted and chosen.

How much are customers willing to delegate to AI?

According to McKinsey, 84% of European consumers now use AI in their everyday lives, and more than a third of them bring it into their shopping decisions specifically. That's a striking number for a technology that barely featured in mainstream shopping habits a couple of years ago.

But look closer at what people actually let AI do, and a more cautious picture appears. Go back to your swimming suit search. You were happy to let the assistant narrow the whole market down to three options and explain its reasoning. Would you be just as happy letting it enter your card details and place the order without checking with you first? Most people, it turns out, wouldn't. McKinsey's research shows European shoppers are comfortable using AI to compare, summarise and recommend, but that comfort drops sharply the moment AI moves from advising to acting.

Researchers describe AI shopping delegation as a curve rather than a switch. At one end sits something as simple as an assistant autofilling your saved cart. At the other sits full autonomy, where an agent transacts on your behalf across a live web of retailers without you checking in at all. Most of us, most of the time, sit somewhere in the middle: happy to be advised, less happy to be replaced.

Where a category falls on that curve depends a lot on what's actually at stake. Restocking laundry detergent is low regret: if the AI gets it slightly wrong, you shrug and reorder. That's exactly why groceries and household staples are where delegation is moving fastest. Buying a luxury handbag is the opposite. It's tied up in identity and taste and how you want to be seen, so people keep the final call for themselves and use AI mostly for research rather than the purchase itself.

From "nike shoes" to "best running shoes for winter"

Something else is changing quietly underneath all of this: the way people search in the first place.

For years, shopping online meant typing a few keywords into a search bar. You knew roughly what you wanted, so you typed the brand and the product, maybe the size, and scrolled through the results. AI chatbots are changing that habit. Instead of typing "nike shoes," people increasingly ask something closer to "what are the best running shoes for winter running on icy surface," and expect a real answer, not a list of links.

That's a meaningfully different question to answer. A search engine only has to match keywords. A chatbot has to actually understand the intent behind the question, weigh up the options and justify a recommendation.

For retailers, this means visibility no longer depends on search engine optimisation alone. Ranking on Google still matters, but it isn't enough if an AI agent is the one doing the comparing and recommending on the customer's behalf. Retailers now also need generative engine optimisation, or GEO. Miss that shift, and a brand can be technically visible online while being functionally invisible to the tool that's increasingly doing the shopping.

Who's building this, and what it means for advertising

Every major tech platform wants a piece of it. Analysts expect the virtual shopping assistant market to pass $8 billion by 2032, and more than half of consumers already say they'd be comfortable letting an AI agent place an order for them. Google is introducing a Universal Commerce Protocol to plug AI agents directly into retailers' catalogues. Perplexity built "Shop Like a Pro" to curate a purchase and check out in one click. Google Lens now lets you photograph a product on the street and instantly compare prices and check nearby stock.

AI agents could fundamentally change advertising. Search ads and retail media exist to catch a shopper's attention at the exact moment they're deciding what to buy, but that's exactly the moment an AI agent skips. Spend will move earlier in the journey instead, away from keyword bidding and toward the kind of brand building, emotional connection and trust that shapes what an AI recommends.

What happens to the physical store?

None of this spells the end of the physical store, but it does mean stores need a clearer reason to exist. Store visits are becoming less frequent but more valuable, splitting into two distinct jobs: convenience and discovery.

Inside those stores, something else is changing too. AI-supported "clienteling" tools are trickling down to mainstream retail, giving staff real-time access to inventory, customer history and next-best suggestions right there on the shop floor. And even though so much of the shopping journey now happens through AI, classic SEO hasn't become irrelevant.

The trust gap: why people still hesitate

If AI shopping is this capable, why isn't everyone using it for everything already? The honest answer is trust, not technology.

Statista's research on AI-assisted shopping found that 52% of consumers worry about being misled by fake or biased content, and 46% are concerned about privacy. Only 6% say they have no concerns at all.

Retailers are behind on this too. As of 2023, fewer than 15% of wholesale and retail companies in the EU had fully implemented AI tools, even though more than 70% were experimenting with them. Businesses are still catching up to a level of AI readiness their customers are already expecting.

What retailers need to do now

Put together, these shifts point to a fairly clear to-do list.

• Product data needs to be clean, current and genuinely machine-readable.
• SEO and GEO need to run side by side.
• Store staff deserve the same AI-powered tools that customer-facing chatbots get.
• Trust, fulfilment reliability and transparent policies matter more than ever.

The funnel isn't disappearing, it's compressing

Come back to that swimsuit search one more time. A few years ago, it would have taken you across a dozen browser tabs, comparison sites and reviews before you settled on one. Today, it can happen in a single conversation, with an AI doing the comparing for you.

By 2030, McKinsey projects that up to $5 trillion in global commerce could be mediated by AI agents. Retailers who treat that as a marketing gimmick are missing the real risk: being quietly filtered out before a customer ever reaches their site or app. The funnel isn't going away. It's just starting somewhere new, in a conversation instead of a search box.