Today’s highest-converting shoppers start their buying journey with AI. Sometimes they know exactly what they want. Other times, they just know what they need: relief from a headache, a warmer coat, a gift for someone impossible to shop for. Either way, they ask an AI assistant to narrow their options, get recommendations, and click through to a storefront with the decision mostly made, ready to buy.
The catch is what happens after the click. A slow site, irrelevant search results, or a generic experience built for no one in particular can turn a near-certain sale into a quick exit. On the flip side, AI-referred shoppers convert 42% better than other traffic, stay on site longer, and spend more per visit than traditional channels, according to Adobe.
In the age of agentic commerce, getting found by AI is only valuable if the experience on the other side is worth the visit.
Speed is a sales strategy
AI-referred shoppers arrive already informed and more engaged than typical traffic: 27% lower bounce rate, 38% longer time on site, 10% more page views per visit, Adobe reports. But in that make-or-break moment, that engagement is fragile, and retailers must meet their intent quickly.
Seven in 10 consumers say page speed directly influences their willingness to buy, with conversions dropping by as much as 17% for each additional second of load time, according to PayPal. The damage escalates the closer shoppers get to checkout and hits hardest on mobile, where 53% of users abandon sites that take longer than three seconds to load, per Google researchers.
Load speed is one culprit, but any friction that slows the buying experience puts the sale at risk. Cart abandonment already runs at a global average of 70.22%, with long checkout flows, forced account creation, and missing trust signals among common causes, per Baymard Institute.
“AI-referred shoppers expect at least the same or better experience when they come to your site,” says Poornima Agoramurthy, principal product manager at Adobe Commerce. “If they get something slower, less relevant, and the experience doesn’t keep up with what they’re getting elsewhere, they lose confidence in the brand very quickly.”
A fast buying experience is now the price of admission.
Search that works the way shoppers think
Once a shopper lands, on-site search is one of the most common places retailers lose them. According to Baymard Institute, 56% of e-commerce sites deliver poor on-site search experiences, built around keyword logic that requires shoppers to think like a product catalog, rather than a person.
The mismatch hits especially hard with AI-referred shoppers. “If customers come from a conversational experience where they’ve had a chance to go back and forth with the AI, and they have to move to classic keyword search on your site, that’s a real turnoff,” cautions Shannon Hane, principal product marketing manager at Adobe Commerce.
Natural-language search closes that gap. Someone searching for “comfortable office chair for long work days” or “waterproof jacket for mountain hiking” shouldn’t have to guess which keywords your product descriptions use.
“Consumers expect to interact using natural language; that’s table stakes now,” says Agoramurthy. “Even if you’re not ready to make your entire storefront conversational-ready yet, AI-powered search is the first step. It responds to the intent and context a shopper is coming with, not just the words they typed.” For AI-referred shoppers especially, search is a potential off-ramp. If the on-site search experience doesn’t match the conversational quality of the AI assistant that sent them there, it’s a reason to leave. Adobe Commerce’s newly released Semantic Search is built to close that gap, delivering the same intent-driven, contextual experience shoppers now expect.
From personalization to purchase, without losing the thread
Most personalization today is backward-looking, built on browsing history, past purchases and demographic segments. That works reasonably well for returning customers. It falls short for AI-referred shoppers, who arrive with something more valuable than behavioral data: declared intent. They already told an AI what they need. The storefront should respond to that signal.
Agoramurthy describes what it feels like when it doesn’t: “Spending 30 minutes interacting and building context with an AI assistant, only to lose it entirely when switching to another site, creates unnecessary friction and frustration.”
Conversational commerce takes that responsiveness further. Consider a first-time marathon runner arriving at a sporting goods site looking for shoes. Rather than clicking through endless product pages, they ask a conversational agent: “What do I need to run a race in 30 days?” The agent recommends the right shoes and athletic wear, provides nutrition advice leading up to race day, and connects them to a local running community — all without leaving the chat interface. “We’re seeing conversational experiences go beyond ‘here’s the best product, buy it,’” says Agoramurthy. “It’s more about building a real relationship with the consumer.”
That kind of experience is the aim of Adobe Brand Concierge, which integrates natively with Adobe Commerce to deliver real-time pricing, inventory, and recommendations within a single conversational interface. “You get an authentic interaction that represents the brand, with the guardrails to make sure it’s enterprise-ready,” says Hane.
It’s rapidly becoming standard practice. Already, 79% of e-commerce brands report that conversational commerce has increased sales, according to eGrow.
The experience is the closer
Converting and retaining AI-referred shoppers hinges on whether the experience on the other side of the click matches the intent that brought them there.
For retailers evaluating solutions, Hane recommends prioritizing four capabilities:
- Personalized recommendations grounded in real-time product data, including price and availability
- Merchant control over brand voice and guardrails so the AI represents the brand authentically
- End-to-end coverage from guided discovery through post-purchase loyalty
- A single platform that unifies data and experience rather than stitching point solutions together
Retailers that build this kind of infrastructure also gain something most don’t have today. “You’ll have access to all the questions and prompts customers are entering,” says Hane. “The LLMs of the world are collecting all that information right now. It’s incredibly valuable for site design, product offerings, and content personalization — and also for understanding how consumer needs and behaviors are evolving to improve how your products align with current customers' needs.”
A shopper who feels genuinely helped has a reason to come back, and a reason to trust the brand the next time an AI assistant makes a recommendation. Without that experience, retailers risk a second kind of invisibility: found by AI, but not worth returning to.