Sixty percent of Germans can imagine using an AI shopping agent. Only 9% would actually let one complete a purchase, including payment. That 51-point gap between curiosity and trust, measured in a Bitkom survey of 1,005 people aged 16 and older, is the single most important number in European agentic commerce right now. It means the largest e-commerce market in Europe is interested in what AI agents can do but fundamentally unwilling to hand over the credit card.
This is not a technology problem. Amazon Rufus, OTTO’s new Gemini-powered assistants, and the Zalando Assistant are all live in Germany today. The infrastructure exists. The consumer trust does not.
The 51-Point Trust Gap: What German Consumers Actually Think
The Bitkom numbers tell a layered story. Among the general population, interest in AI-assisted shopping is real. People want help finding products, comparing prices, and filtering reviews. But the moment “AI agent” shifts from “advisor” to “buyer,” adoption collapses.
The breakdown by age makes it sharper. Among Germans aged 16 to 29, 43% would let an AI handle purchases autonomously. For the overall population, it is 9%. That is not a generational gradient. That is a cliff. And for retailers planning agentic commerce rollouts, it means the majority of their current paying customer base actively rejects the core premise.
Germany is not alone in this skepticism, but the intensity is unusual. In the US, Contentsquare found that 30% of consumers would let an AI agent complete a purchase. ChannelEngine’s 2026 report, covering 4,500 shoppers across the US, UK, France, Germany, and the Netherlands, found only 17% feel comfortable with AI-completed purchases overall. Germany consistently sits at the bottom of that range.
Why the Gap Is Wider in Germany
Three factors explain why German consumers are more resistant than their US or UK counterparts.
Privacy as identity, not preference. In the US, data sharing is a transaction: you trade personal information for convenience. In Germany, Datenschutz is closer to a constitutional value. The DSGVO is enforced more aggressively in Germany than anywhere else in Europe, and consumers have internalized those norms. When IBM’s Institute for Business Value found that 83% of consumers globally worry about privacy, data misuse, and unwanted marketing from AI shopping agents, you can assume that number runs higher in Germany.
The Verbraucherschutz reflex. Consumer protection is not just a policy in Germany; it is an institution. The Verbraucherzentralen (consumer advice centers) actively investigate and litigate against companies that overstep on data use. They already fought Meta on AI training data. AI shopping agents that store payment credentials, track purchase preferences, and act autonomously are a natural next target.
No cultural muscle for “move fast.” Germany’s compliance culture treats ambiguity as a stop sign, not a yellow light. When regulations around AI agents are unclear, German companies and consumers do not experiment cautiously. They wait. This is the same pattern documented in Germany’s broader AI adoption bottleneck, where 62% of workers felt ready for AI but a third had never used it professionally.
Who Is Deploying Anyway: OTTO, Zalando, and Amazon in the German Market
Despite consumer resistance, Germany’s biggest retailers are pushing forward with AI agents. The strategy is not to convince skeptics. It is to start with low-risk use cases and expand from there.
OTTO’s Two-Agent Approach
On March 5, 2026, OTTO launched two AI assistants built with Google Gemini. One handles product discovery: helping shoppers find items, compare options, and narrow choices. The other handles customer service: tracking orders, processing returns, answering FAQ-style questions.
Neither agent buys anything. OTTO’s approach respects the trust gap by keeping AI in an advisory role. The shopper always makes the final purchase decision. Internally, OTTO also runs ogGPT, an AI assistant used by over 7,500 employees monthly for operational tasks. The company is building AI competence internally before pushing it externally.
Zalando’s Cautious Growth
Zalando’s AI-powered shopping assistant reached 6 million users in 2025, a 4x increase year-over-year. The assistant recommends outfits, helps with sizing, and surfaces products based on conversational input. It does not handle checkout.
Zalando also endorsed Google’s Universal Commerce Protocol (UCP), an open-source standard for agentic commerce that lets AI agents interact with product catalogs, inventory, and pricing through a standardized API. But UCP is currently US-only, with global expansion planned. Zalando’s endorsement is a bet on the infrastructure, not a near-term German deployment.
Amazon Rufus in Germany
Amazon Rufus has been available in Germany since late 2024 and is on pace for a $10 billion annualized sales impact globally. In Germany specifically, Rufus operates primarily as a search and recommendation tool. Amazon’s newer “Buy for Me” feature, which lets the AI complete purchases on third-party websites, triggered a backlash from 180+ brands and has not been aggressively marketed in Germany.
Why OpenAI’s Checkout Failure Matters for Germany
OpenAI’s Instant Checkout, a feature that let ChatGPT users buy products without leaving the conversation, lasted about four months. Only around 30 Shopify merchants were onboarded. Users browsed and compared enthusiastically but almost nobody completed a purchase inside the chatbot.
The failure validates what German consumers already believe: people want AI to help them shop, not to shop for them. The use pattern was clear across all markets. Consumers used ChatGPT as a research tool, asked questions, compared specs, read reviews, then left to buy on the retailer’s own site where they had existing accounts, saved payment methods, and loyalty programs.
For the German market, this is doubly significant. If American consumers with higher baseline trust in AI and less privacy consciousness still refused to buy through a chatbot, the prospect of German consumers doing so is even more remote.
The Perplexity Lawsuit Sets a Precedent
Amazon won a preliminary injunction blocking Perplexity’s AI shopping agent, Comet, from operating on Amazon in March 2026. The court found that Comet accessed Amazon’s platform with user permission but without Amazon’s authorization. This case has direct implications for the European market, where the Digital Markets Act explicitly promotes interoperability but has not been tested against autonomous AI agents acting on a user’s behalf.
For German consumers, the legal uncertainty adds another reason to hesitate. If even the companies building these agents cannot agree on who is authorized to do what, why would a consumer trust the process?
The Fraud Vector Nobody Is Talking About
Consumer skepticism in Germany may be justified by a threat that most agentic commerce coverage ignores entirely. Visa’s 2026 threat assessment documented a 450% increase in dark web mentions of AI agent exploitation techniques. The specific risks include:
- Fake merchants optimized for AI agents. Unlike human shoppers, AI agents evaluate structured data: prices, reviews, ratings. Fraudsters are building storefronts that look perfect to an algorithm but deliver nothing.
- Agent credential theft. If an AI agent stores your payment details and purchase preferences, compromising that agent gives an attacker everything they need. Not just one transaction but your entire purchasing profile.
- Prompt injection attacks on shopping agents. Malicious product descriptions or reviews can contain instructions that manipulate an AI agent’s behavior, redirecting purchases or exfiltrating user data.
Visa introduced the Trusted Agent Protocol (TAP) to address some of these risks and completed secure AI transactions with 10+ partners. But the infrastructure is early, and fraud prevention for agentic commerce remains largely theoretical.
For a market where 83% of consumers already worry about AI and privacy, adding “your AI agent might get hacked and buy things you never wanted” to the conversation is not going to accelerate adoption.
What This Means for DACH E-Commerce Strategy
The data points in a clear direction for retailers operating in Germany, Austria, and Switzerland.
Start advisory, not autonomous. OTTO’s two-agent model is the template. Let AI handle product discovery, comparisons, and customer service. Keep checkout under human control. The gap between “interesting” and “trusted” is too wide to skip.
Lead with transparency. German consumers respond to explainability. Show them what the AI recommended and why. Let them override every suggestion. Build trust through visible control, not hidden convenience.
Solve the liability question before scaling. The European Business Magazine’s analysis asks the right question: when an AI agent makes an unauthorized or harmful purchase, who pays? No EU regulation currently answers this. The Digital Fairness Act, expected in Q4 2026, might address it. Until then, retailers deploying autonomous purchasing in the EU carry unquantified legal risk.
Watch the generational divide. The 43% acceptance rate among 16-to-29-year-olds is not nothing. These are tomorrow’s primary spenders. Building AI shopping experiences for this cohort while respecting the broader market’s boundaries is the strategic play.
Bain projects agentic commerce at $300 to $500 billion in the US alone by 2030. McKinsey puts the global figure at $3 to $5 trillion. Those numbers assume consumer adoption follows technology availability. In Germany, the data says it will not. The market that buys more online than any other in Europe is also the market least willing to let a machine do the buying. Any DACH commerce strategy that ignores this gap is planning for a customer that does not yet exist.
Frequently Asked Questions
Why are German consumers more skeptical of AI shopping agents than other markets?
Germany’s deep privacy culture (rooted in DSGVO enforcement), strong consumer protection institutions like the Verbraucherzentralen, and a compliance-first organizational culture that treats regulatory ambiguity as a stop sign all contribute. Only 9% of Germans would let an AI agent complete a purchase autonomously, compared to 30% in the US.
Which German retailers use AI shopping agents?
OTTO launched two Google Gemini-powered AI assistants in March 2026 for product discovery and customer service. Zalando’s AI shopping assistant has 6 million users. Amazon Rufus is available in Germany for search and recommendations. None of these agents handle autonomous checkout in the German market.
Is it safe to let an AI agent buy things for you?
Current risks include fake merchants optimized to fool AI agents, credential theft from agent systems storing payment details, and prompt injection attacks through malicious product listings. Visa documented a 450% increase in dark web discussions about AI agent exploitation. Security infrastructure like Visa’s Trusted Agent Protocol is emerging but still early.
What is agentic commerce and how does it differ from normal online shopping?
Agentic commerce means AI agents autonomously browse, compare, and purchase products on behalf of consumers. Unlike traditional e-commerce where humans click “buy,” the AI handles the entire transaction. Bain projects this market at $300 to $500 billion in the US by 2030, but consumer adoption, especially in Germany, remains a major barrier.
Who is liable when an AI shopping agent makes a bad purchase?
No EU regulation currently answers this question. The Digital Fairness Act, expected in Q4 2026, may address AI agent liability in consumer transactions. Until then, it is unclear whether the consumer, the retailer, the AI platform, or the agent developer bears responsibility for unauthorized or harmful AI-driven purchases.