A few years ago, "AI in e-commerce" meant a recommendation widget at the bottom of a product page. Today, it means something far bigger: software that can negotiate a price, reorder stock before you notice it's low, resolve a customer's refund without a human ever reading the ticket, and — increasingly — complete the purchase itself, on a shopper's behalf, without either party touching a checkout button.
This isn't a UK story, a US story, or an Asia-Pacific story. It's happening everywhere at once, at a pace that's outstripping almost every previous wave of digital commerce, including mobile and social shopping. We pulled together what's currently being reported across markets — from McKinsey and Gartner to DHL's 2026 global trends survey of 29,000 consumers and 5,800 businesses spanning Europe, North America, Asia-Pacific, Latin America, the Middle East and Sub-Saharan Africa — to map out where AI in e-commerce actually stands today, and where the real opportunity for growing businesses lies.
The numbers behind the shift
The headline figure is simple: AI has moved from "interesting pilot" to "operational infrastructure" in about 2 years. McKinsey now finds that 78% of organisations use AI in at least one business function, up from 55% in 2023, and within retail and consumer goods specifically, that figure rises to 89%. The global AI-in-e-commerce market itself is on a steep curve — estimates vary by methodology, but most trackers put it somewhere between $9 billion and $10 billion in 2026, climbing toward $50–65 billion by the early 2030s.
What's more telling than the market-size figures is how fast consumer behaviour has shifted. In 2024, just 38% of consumers had used generative AI at any point in their shopping journey. By 2025, that had risen to 51%. Among Gen Z shoppers specifically, around 6 in 10 now use AI tools to help with a purchase, and nearly half of Gen Z and Millennial shoppers say they use AI platforms daily. During the most recent global holiday shopping season, generative AI and AI agents are estimated to have driven roughly $262 billion in retail revenue worldwide — about one-fifth of total holiday sales — and shoppers arriving via AI sources convert around 31% more often and spend 45% more time on retailer sites than those arriving through traditional search.
Regionally, the picture is genuinely global rather than concentrated in one market. North America still leads on AI investment and infrastructure. But e-commerce companies across Asia-Pacific are embedding AI just as aggressively — China and India are ahead of the curve on adoption, with Southeast Asian retailers expanding fast behind them, driven by digital payments growth and booming social commerce. Europe is moving more cautiously on consumer-facing autonomy but is pushing hard on AI-driven logistics and supply chain efficiency. And DHL's 2026 survey — covering the Middle East and Sub-Saharan Africa as well as the more commonly discussed markets — found that nearly a third of shoppers globally, rising to over a third of Gen Z and Millennials, say they'd be comfortable handing control of their shopping over to AI entirely.
It helps to see this as an evolution rather than a single leap. The first wave of e-commerce AI — still in use almost everywhere — was about prediction: "customers who bought this also bought that," dynamic search ranking, basic chatbots answering FAQs from a script. Useful, but fundamentally reactive.
The second wave, generative AI, changed the content layer: product descriptions, marketing copy, customer service responses, and, increasingly, the shopping interface itself, as consumers started asking ChatGPT, Gemini, or Perplexity to help them research and compare products rather than typing into a search box.
The wave now underway is different in kind, not just degree. It's agentic — systems that don't just predict or generate, but reason through multi-step problems and act across connected business systems with limited human supervision. Google Cloud's VP of Global Solutions, Carrie Tharp, has described this as AI evolving "from a passive tool that offers prediction, to active, autonomous resources that can execute complex, multi-step, prescriptive actions across every consumer and operational touchpoint." That's the shift that matters for any business trying to scale: the software is no longer just advising your team — it's starting to do the work itself.
An AI agent, in this context, is software that can observe data from multiple sources, reason about their meaning, and then take action across connected systems — without a human approving every individual step. That's meaningfully different from a chatbot, which can only respond when prompted and typically follows a fixed script.
Across global retailers and growing brands alike, AI agents are now doing real operational work in several areas:
Customer service and support. This is the most mature use case. Klarna's AI assistant reportedly handled roughly two-thirds of customer service chats in its first month of full deployment — around 2.3 million conversations — doing the equivalent work of 700 full-time agents and contributing to a 25% drop in repeat inquiries. More broadly, industry data suggests that 80% of customer service organisations are now integrating generative AI into support workflows, and businesses report 40–60% reductions in support costs when agents are deployed effectively.
Inventory and demand forecasting. Rather than reacting to a stockout after it happens, agents now monitor sell-through in real time, forecast demand at the SKU and location level, and automatically trigger replenishment. One global luxury fashion brand using an AI-driven inventory platform reported a 10% reduction in inventory costs alongside a 12% increase in SKU availability after implementation.
Dynamic pricing. Pricing agents adjust in real time based on competitor moves, inventory age, and demand signals — for example, automatically raising the price of a product when a competitor goes out of stock on the same item, while staying within the margin guardrails set by the business.
Content at scale. Writing product descriptions for an entire catalogue manually can take 25–30 hours for just 100 SKUs. AI content agents now do the same job in minutes, while also optimising for how AI search engines and shopping agents — not just human readers — parse product information.
Fraud and risk. Agentic fraud detection moves the function from reactive (catching fraud after the fact) to proactive, flagging suspicious order patterns before a transaction completes — increasingly important as AI-driven shopping itself opens new fraud vectors that financial institutions are actively bracing for.
Logistics and fulfilment. Agents are increasingly monitoring supplier data, delivery timelines and warehouse systems together, flagging a likely delay before it ever reaches the customer rather than after a complaint comes in.
The most dramatic development of the last twelve months is what the industry is now calling agentic commerce — AI doing the buying, not just the recommending. The infrastructure for this has moved from theoretical to operational extremely quickly. Stripe's Agentic Commerce Suite, launched in December 2025, already powers checkout for major brands across fashion, accessories and homeware. Google launched a Universal Commerce Protocol in early 2026 with backing from Shopify, Etsy, Wayfair and Target, and endorsement from payment networks including Visa, Mastercard and American Express. Amazon's Rufus assistant is now serving an estimated 300 million users and is credited with billions of dollars in incremental sales; shoppers who use it convert at notably higher rates than those who don't.
Not every brand has navigated this transition smoothly, which is itself an instructive data point. Walmart reportedly found that purchases completed directly in a chat interface were three times worse than when it redirected shoppers to its own website — so it changed its approach, building its own AI assistant and integrating it into platforms like ChatGPT rather than ceding the entire customer relationship. That nuance matters: agentic commerce isn't simply "let the AI take over." It's a new layer that businesses need to shape, not passively accept, but actively.
Forecasts on how big this gets vary, which is itself a sign of how early we are. Bain projects that the US agentic commerce market could reach $300–500 billion by 2030, accounting for 15–25% of total e-commerce sales. McKinsey estimates the broader global opportunity at $3–5 trillion. Gartner expects 40% of enterprise applications to embed AI agents by the end of 2026. Whatever the precise number turns out to be, the direction is unambiguous and global — from US retail giants to Asia-Pacific marketplaces to European fashion and homeware brands now integrating with these same protocols.
It's tempting to frame all of this purely as a customer-experience story. The more important story, especially for ambitious or growing businesses, is about scaling operations without scaling cost and headcount at the same rate.
This is where AI agents earn their keep. A growing e-commerce business traditionally hits a wall: more orders mean more support tickets, more inventory complexity, more content to produce, more pricing decisions to make — all of which historically meant hiring more people, in roughly the same proportion as growth. Agentic AI breaks that link. The U.S. Chamber of Commerce reports that small-business AI adoption has roughly doubled since 2023, with high-tech small businesses reporting markedly stronger sales and profit growth than low-tech competitors. Separately, businesses deploying AI agents across e-commerce operations report revenue 30% higher than non-adopting competitors, alongside meaningful reductions in operational costs — gains that stem from handling more volume and complexity without proportional headcount growth.
Crucially, this opportunity isn't reserved for retail giants with in-house AI teams. The infrastructure has matured to the point where mid-sized and growing brands can deploy purpose-built agents — for support, inventory, content, and pricing — incrementally, starting with whichever workflow is causing the most pain and expanding from there. One industry voice put it well: the real divide in the next few years won't be big businesses versus small ones, so much as businesses with clean data and modern systems versus those without, meaning a well-run, digitally mature smaller brand can move just as fast as a household name.
None of this comes for free, and a fair piece on this topic should also give equal attention to where AI in e-commerce is genuinely struggling right now.
The most consistent finding across recent research is an execution gap. Only around a quarter of companies report having developed the organisational capability to generate measurable value from AI, even as the large majority are experimenting with it — meaning roughly three-quarters are still working out how to turn pilots into ROI. Supply chain leaders are particularly candid about this: while around three-quarters identify AI as their primary strategic driver, fewer than a third currently have the underlying data infrastructure to act on it. A lot of that comes down to unglamorous but essential groundwork — fragmented systems, batch-processed data updates that lag reality by hours, and product catalogues that aren't structured in a way that AI agents can actually parse.
That last point deserves emphasis, because it's becoming the dividing line for AI-driven discovery specifically. AI agents are far less forgiving of messy data than human shoppers ever were: where a person might tolerate a vague product description, an agent is more likely to skip a listing with incomplete or inconsistent attributes. Pages with properly structured product data are already being cited multiple times more often in AI-generated shopping answers and overviews than pages without it. In practice, this means the technical groundwork — clean catalogue data, schema markup, real-time inventory feeds, API-accessible systems — is no longer a back-office nicety. It's the foundation that determines whether AI agents can even see your products.
There are also legitimate concerns to manage rather than dismiss: data security and privacy (cited by 44% of CEOs as their top AI concern), the risk of fraud growing alongside agent-mediated transactions, and the question of how much decision-making to hand over without human oversight, particularly for refunds, pricing exceptions and anything touching customer trust directly. The businesses winning right now aren't the ones deploying AI everywhere at once — they're the ones starting with well-scoped, high-volume, low-ambiguity workflows (order status, routine returns, replenishment) and keeping a human in the loop for the edge cases and high-stakes decisions, expanding agent autonomy only as confidence and data quality improve.
For a business looking at all of this and wondering where actually to begin, the pattern across the research is consistent:
Start by getting product, inventory and customer data into a clean, structured, machine-readable state — this is the unglamorous prerequisite that almost every other AI capability depends on. Pick one high-friction workflow to automate first, rather than trying to deploy AI everywhere simultaneously; customer support and inventory forecasting tend to offer the fastest, most measurable returns. Build in human oversight for anything involving money, policy exceptions or brand-sensitive judgment calls, and loosen that oversight gradually as the system proves itself. And treat your product data the way you'd treat your website's SEO a decade ago — because being legible to AI shopping agents is rapidly becoming as important as being legible to a search engine, regardless of which country or platform your customers are shopping from.
The throughline across every market, every report and every region surveyed is the same: AI in e-commerce has stopped being a competitive nice-to-have and become foundational infrastructure — the same way mobile-responsive websites and secure payment gateways once were. The businesses pulling ahead aren't necessarily the ones with the biggest budgets; they're the ones treating AI agents as a genuine operational layer — handling support, inventory, pricing and content at scale — while keeping the data foundations and human judgement that make that layer trustworthy.
That combination — solid e-commerce architecture, clean data, and AI built in as infrastructure rather than bolted on as a gimmick — is exactly the intersection where SAM AI Solutions works, building e-commerce platforms designed from the ground up to support AI-driven personalisation, automation and the agentic systems that are quickly becoming the new normal across global retail.
Find out more → samaisolutions.co.uk
