Most e-commerce brands are still optimising for a customer journey that fewer and fewer people are actually taking. The assumption of browse, search, compare, click, buy made sense for a long time. That is changing, and faster than most brands have accounted for.
A growing share of your customers are now starting that process somewhere else entirely: inside an AI. They’re asking ChatGPT what to buy for a wedding anniversary, letting Perplexity compare luxury candles and asking Gemini to find the best-reviewed gifts under £100. For most shoppers today, AI is still playing an advisory role. It surfaces options, compares products and shortlists choices before the customer makes the final decision. But the share of that decision-making happening inside AI tools, rather than on your store, is growing fast.
This isn’t a prediction. It’s already happening, and most brands have two significant blind spots at once.
The discovery problem
AI shopping agents don’t browse the way humans do. They don’t scroll category pages or respond to homepage banners. They pull structured data, parse product feeds and make decisions based on what they can actually read and verify. If your product information is incomplete, inconsistent or missing key attributes, agents may not flag it. They move on, and your competitor gets surfaced instead.
The scale of this shift is already showing up in traffic data. AI-referred traffic to retail sites grew 693.4% across the 2025 holiday season, according to Adobe Analytics, though it’s worth noting this is from a low base. eMarketer forecasts AI platforms will drive $144 billion in e-commerce sales by 2029, nearly 9% of all US retail e-commerce. eMarketer also projects that 20% of all online transactions could be initiated by AI agents by the end of this year, a figure echoed by Gartner, though “initiated” covers a wide range of behaviour, from an AI recommendation that leads to a click through to fully autonomous purchasing, which remains rare.
A joint study by IBM’s Institute for Business Value and the National Retail Federation, surveying 18,000 consumers across 23 countries, found that 45% are already using AI somewhere in their buying journey. Adyen’s 2026 UK Retail Report found that 44% of UK shoppers are now open to having AI handle the entire shopping process, including the final purchase, once their preferences and budget are set. This is a shift still in progress, but the direction is clear.
A meaningful and fast-growing portion of your addressable market is making purchase decisions in an environment where your product content needs to work harder than ever.
The brands currently performing well in these environments aren’t necessarily the biggest or the best-known. They’re the ones whose data is structured, complete and consistently maintained across every feed and platform. Missing dimensions, vague product descriptions, absent review data and unstructured specifications are no longer just bad for traditional SEO. As AI search becomes a more significant part of how products are found and evaluated, the stakes attached to data quality have risen considerably.
Maintaining that consistency across feeds, platforms and a constantly changing product catalogue is demanding but increasingly essential.
The experience problem
The expectations of customers who still land directly on your store are also shifting.
People who regularly use AI tools in their daily lives, for research, for recommendations, for planning, are increasingly accustomed to experiences that feel relevant to them specifically. Not just “personalised” in the sense of a name in an email subject line, but genuinely tailored. Recommendations that reflect their actual preferences, interfaces that surface what they’re likely to want before they go looking for it.
When those customers arrive on an e-commerce site that doesn’t match that experience, the gap is noticeable. According to Adobe Analytics, visitors referred from AI platforms convert at a rate 31% higher than average and generate 254% more revenue per visit, suggesting that when the experience matches the expectation AI has already set, shoppers are significantly more ready to buy. Customers who don’t get that experience are increasingly likely to leave without converting, having been conditioned by AI-assisted experiences to expect something better.
This isn’t a feature problem. It’s a foundations problem.
Brands that are getting this right aren’t deploying one big solution. They know what a returning customer looks like, and they can adjust what products are shown and in what order.
The visibility gap
What makes this difficult is that the signals don’t always show up where you’d expect them to. AI referral traffic, Google AI Overviews and brand visibility across recommendation platforms are increasingly trackable, but the picture is rarely complete, and interpreting it requires a different kind of attention than reading a standard analytics dashboard.
The earlier part of the funnel is harder still. When AI tools evaluate products during the recommendation phase, that process doesn’t show up in your usual reports. The gap tends to surface indirectly, in conversion rates that don’t quite add up or revenue that’s hard to account for directly.
As McKinsey notes, most brands are still running personalisation efforts as manual, one-off experiments rather than measuring them in any integrated way, which means the effect often gets absorbed into overall performance numbers without a clear cause.
What ties all of this together is the quality of the foundations underneath. Structured product data, clean feeds, solid SEO and proper analytics infrastructure are what influence visibility in AI search and traditional search alike. The brands least likely to face these problems are those who’ve already got that groundwork in place.
These things compound over time. AI systems learn which products and brands to surface based on data quality and engagement signals. Stores that aren’t visible or competitive in AI-referred channels today will become progressively less so. The window to build the right foundations is narrowing.
It is also where the right expertise makes the most difference.
AI hasn’t changed what good e-commerce looks like at its core: relevant products, trustworthy information, a smooth path to purchase.
Emrise works with e-commerce brands on the foundations that influence both search visibility and AI recommendation, from structured product data and technical SEO to analytics and measurement. Get in touch to find out where your store stands.
