Retailers have spent years refining the digital shopping journey, optimising search, polishing product pages and investing heavily in performance marketing to capture consumer attention at the point of intent. But increasingly, that intent is being shaped before a shopper ever reaches a retailer’s website.

Artificial intelligence is fast becoming the new shop window to retail discovery. Consumers are turning to tools such as ChatGPT, Gemini and other AI-powered assistants to compare products, weigh up prices, summarise reviews and narrow down their options long before they click through to buy.

In other words, the early and influential stages of the customer journey are moving away from traditional search and into conversational interfaces.

The rammifications here for retailers are huge. It changes where decisions are being made, and who or what is influencing them. For retailers, the challenge is increasingly about ensuring products are visible, credible and competitive inside AI-generated recommendations.

The scale of the shift is becoming harder to ignore. McKinsey research shows that 84 per cent of consumers across the UK, France and Germany now use AI tools in their daily lives, while 38 per cent are already using them to research products or inform purchasing decisions.

Crucially, this usage is concentrated at the top of the funnel, where shoppers are exploring categories, comparing brands and deciding what’s worth their time.

According to McKinsey’s data, 63 per cent of European consumers who use AI in shopping are using it to compare brands, models, prices and reviews. Another 55 per cent are using it to learn about categories or products, while 46 per cent are turning to AI for inspiration and discovery.

This suggests AI isn’t simply becoming another channel layered on top of ecommerce, but instead becoming a decision-support engine in its own right, helping consumers condense a once fragmented process into a single experience.

That behavioural change is already feeding through into traffic and conversion data. Adobe Analytics found that traffic from generative AI tools to retail websites surged by an eye watering 693 per cent year-on-year during the 2025 holiday season.

Retail saw the sharpest growth of any sector, ahead of travel, financial services, technology and media. In November alone, AI-driven traffic to retail sites jumped by 769 per cent, followed by another 673 per cent rise in December.

More importantly, this isn’t low-quality ‘curiosity’ traffic. These shoppers are converting. Adobe found that shoppers arriving from AI assistants converted 31 per cent more often than those from other traffic sources during the holiday period.

Revenue per visit from AI-driven traffic was up by 254 per cent, while those visitors also spent 45 per cent longer on site and viewed 13 per cent more pages per visit. Bounce rates were also markedly lower, suggesting that shoppers arriving from AI tools weren’t just browsing more efficiently, but landing with a much clearer sense of what they wanted.

That points to a significant shift in how product discovery is working in practice.

Consumers aren’t doing the heavy lifting themselves across multiple websites. Instead, they’re increasingly outsourcing that work to AI, asking it to compare options, explain trade-offs and surface the most relevant choices. By the time they arrive on a retailer’s website, the shortlist may already have been made.

This is where trust becomes critical. AI may be growing in influence, but consumers aren’t embracing it blindly. New research from Criteo’s Consumer Shopping Index suggests that UK shoppers are willing to rely on AI in the purchase journey, but only when they believe it is acting impartially.

The study found that 55 per cent of UK consumers trust AI most when it shows side-by-side prices and availability across multiple retailers. Meanwhile, 51 per cent say they would rely on AI through a multi-retailer platform, compared with just 37 per cent on a single retailer site.

That gap suggests that consumers are more comfortable with AI when it behaves like a neutral adviser, rather than a branded sales assistant. The appeal lies not simply in speed or convenience, but in the sense that the recommendation is broader, more balanced and less obviously self-serving.

This creates a more complex competitive landscape for retailers. In a world of AI-mediated discovery, the challenge is to be seen and selected above the crowd.

If an AI assistant is narrowing a field of hundreds of products to just three recommendations, what determines whether a brand makes the cut? Price will matter, of course, but so will availability, reviews, relevance, data quality and the ability to stand up to comparison.

The implications go to the heart of how retailers present themselves in digital environments. If AI becomes the layer through which products are discovered, then product information has to work not only for people, but for machines.

Structured data, accurate metadata, consistent taxonomy and trustworthy third-party signals will all become more important in determining whether a product is surfaced and recommended.

Some are already describing this as the rise of ‘agent engine optimisation’,  a natural evolution of search thinking for a world in which discovery is increasingly mediated by AI agents rather than search engines alone.

The phrase may be new, but the principle is straightforward. If a product cannot be easily interpreted, compared and justified by AI, it risks becoming invisible at the very moment consideration begins.

This shift isn’t confined to one demographic or one kind of shopper. Retail Economics research shows that AI is moving firmly into the mainstream. Globally, 78 per cent of consumers say they have used AI tools such as ChatGPT in the past 12 months, rising to 93 per cent among under-35s.

In the UK, almost half of adults under 45 are already using AI for ecommerce tasks such as product research, price comparison and exploring delivery options.

Chat-based AI platforms are also becoming an important commerce channel in their own right. The Retail Economics data suggests these platforms are now generating 50.2 million shopping-intent visits in the UK every month, putting them alongside some of the most important digital entry points in retail.

Consumers aren’t all approaching AI in the same way, however. The same research identifies four distinct shopper personas emerging around AI.

At one end are ‘AI delegators’, who make up 17 per cent of shoppers and are relatively comfortable allowing AI to discover, compare and even buy on their behalf. Then there are ‘AI collaborators’, who account for 30 per cent and see AI as a trusted co-shopper rather than a substitute decision-maker.

Another 30 per cent fall into the ‘AI selectors’ group, using AI more occasionally for reassurance or research, while 23 per cent remain sceptical and continue to prioritise familiarity and control.

That spread points to where adoption is heading next. Consumers may be increasingly open to AI helping them decide, but they’re still more cautious about AI acting independently.

McKinsey’s findings reinforce this, showing that trust is strongest when AI supports judgement by comparing options, summarising reviews or highlighting trade-offs, but drops off as tasks move closer to execution. In other words, consumers are willing to delegate thought before they are willing to delegate control.

Even so, Retail Economics found that 30 per cent of UK adults are already open to AI acting as a personal shopping agent that could recommend products, check delivery and returns options and, in some cases, complete purchases once preferences are known.

By 2030, 48 per cent of shoppers expect AI to act as a helpful assistant throughout the shopping journey, while 25 per cent expect it to evolve into a trusted co-shopper that automates some decisions.

McKinsey goes even further, estimating that agentic commerce could influence between $3 trillion and $5 trillion in global spending by 2030.

That figure may still feel ambitious, but the consumer behaviours underpinning it are already visible. AI is becoming embedded at the point where preferences are shaped, options are narrowed and purchase intent is formed.

For retailers, the old assumption was that discovery began on a search engine, a marketplace or a brand’s own website. Increasingly, that’s simply no longer the case.

Discovery is starting in AI interfaces that compare options instantly, collapse research time and frame buying decisions before a shopper enters a traditional ecommerce environment.

That means retailers can’t think only about how to convert visitors once they arrive. They must also think about how to remain visible before that visit happens at all.

Click here to sign up to Retail Gazette‘s free daily email newsletter