The Maze: AI shopping has a delegation problem hiding inside an adoption story. Consumers say yes to help, comparison, and price checks. They are much less ready to let an assistant take over the full purchase journey. idealo's six-market Kantar survey shows the split clearly: AI can earn the research layer before it earns the checkout layer.
Guidance is already acceptable at European scale. In idealo's survey, at least 70% of respondents in every measured market say an AI assistant would be helpful if it explains products, compares options, and helps find the best price. Spain leads at 80%, Italy follows at 78%, the UK and Austria sit at 75%, Germany reaches 73%, and France still lands at 70%. That is not a niche use case. It is mainstream appetite for AI as a smarter comparison layer.
Checkout delegation is where trust gets expensive. Approval falls hard when the assistant moves from helping to taking over the entire purchase process. Spain drops to 48%, Italy to 43%, the UK to 41%, Austria to 38%, Germany to 37%, and France to 32%. Germany's 36-point gap is the cleanest lesson: shoppers may accept AI as an analyst, but not yet as the person holding the wallet. That difference matters for every agentic-commerce pitch built around autonomous ordering.
The German data explains why the gap exists. Germany is not anti-AI. 63% of consumers have consciously used AI applications or AI-based features, 68% use them at least weekly, and 41% have already used AI support while shopping. The resistance sits in reliability. The same press release says 68% worry about false or misleading information and 65% worry about data protection. Those are not UX problems. They are proof, liability, and data-quality problems.
Price comparison is the natural wedge. In Germany, price comparison is the top AI-shopping use case at 48%, ahead of product search at 42% and purchase recommendations at 38%. That gives comparison platforms a credible role in the AI commerce stack. The assistant can sound clever, but the transaction needs current prices, real merchant data, delivery details, returns rules, and trustworthy offers. Without that layer, the assistant becomes a confident brochure with a checkout button attached.
Trust builders are boring, which is the point. German respondents name clear labeling of AI-generated content, accurate and reliable results, and strong personal-data protection as the top ways to increase trust. That is not a call for more personality in the assistant. It is a call for source visibility, data freshness, and verifiable recommendations. Jörn Rehse's LinkedIn framing lands here: AI commerce will not be decided only by model quality. It will be decided by whether merchants and platforms can make the result auditable enough to act on.
Why it matters: Agentic commerce is often sold as a single leap from search to autonomous purchase. idealo's evidence suggests it will arrive in stages. First, AI earns the right to explain and compare. Then it earns permission to recommend. Only after that does it get to transact. The winners may not be the flashiest assistants. They may be the data layers that make an assistant safe enough to trust with money.
Sources: idealo press release | Jörn Rehse LinkedIn post
