Author: Wojciech Milewski, Board Brand Advisor, DF4R Date: 8 October 2026 Series: AI / GEO / AEO
---
Lead
AI has become a permanent part of the shopping journey — 82% of the 28,000 consumers surveyed across 17 countries by VML confirm it (Future Shopper 2026, 2.66 million data points). But using a tool is not the same as trusting its output. Most shoppers rely on AI while actively protecting themselves against it — checking, filtering, and limiting its role in the final purchase decision. That gap is now the central challenge for brands building AI-driven shopping experiences.
---
1. The numbers behind the distrust
VML's Future Shopper 2026 is one of the largest global studies of shopping behaviour ever conducted — 28,000 consumers across 17 countries, 2.66 million data points. Its headline number: 58% cross-check AI product recommendations with other sources before acting. That is not the voice of sceptics. It is the voice of the majority.
Other data points confirm the same pattern. 49% say AI-generated product imagery reduces their trust in a brand. 48% skip content they believe was created using AI. 53% worry AI shopping tools are commercially influenced by advertisers, and 55% say sponsored search results make it harder to find the best products. Altogether, 33% would not allow AI to make purchases on their behalf.
The Gemius "E-commerce in Poland 2026" report adds a local counterpoint: 71% of Poles do not want AI to make autonomous purchase decisions. For larger purchases, resistance climbs to 76%. Poland is not an outlier — it is an early warning for the whole of Europe.
2. The confidence gap: technology ready, customers not yet
The vendor ecosystem — AI shopping agents, autonomous bots, GEO/AEO platforms — is growing exponentially. Every week brings new product launches, hundreds of millions in funding, and deeper integrations into search engines and marketplaces.
Yet the consumer message is unequivocal: help me, but don't replace me. Economic pressure (VML reports that nearly 6 in 10 consumers feel greater financial anxiety than ever) does not push shoppers toward handing control to a machine — it pushes them to check everything more carefully. And the growing perception that AI tools are commercially influenced (53%) reinforces the instinct to verify.
This creates what I call a confidence gap: agentic commerce exists on paper and works technically — but society has not granted it trust. Deploying an agent is not a technological challenge. It is a trust challenge.
3. What this means for retailers and brand managers
If 58% of your customers do not trust your AI recommendation, a better model will not fix it. Fix it with:
1. Transparency. Customers need to know when a recommendation comes from AI and what data it is based on. Hiding AI use is a direct route to eroded trust (49% cite this effect explicitly).
2. Verifiability. Give customers the tools to cross-check — comparisons, full specifications, reviews, real-time stock data. Do not fight the instinct to verify; design the process so they come out convinced, not confused.
3. Control. A customer who can approve, modify or reject an agent's action has exponentially higher trust than one who receives a take-it-or-leave-it decision. Let AI offer, but never let it finalise without consent.
VML's Brent Taylor, global CEO of WPP Commerce, puts it plainly: "Long-term loyalty will need to earn consumer confidence at every stage of the customer journey through transparency, consistency, and exceptional customer experience."
4. Three priorities for Q4 2026
- Trust audit of AI channels. Check whether customers know where and how AI participates in their journey, and whether that presence builds or erodes brand trust.
- The human-in-the-loop rule. Any AI feature that completes a purchase without explicit customer consent risks rejection from the third of shoppers who flatly won't accept autonomous buying.
- Structured, verifiable data. Make AI recommendations easy to check and cite, rather than relying on blind customer trust.
Conclusions
Brands rolling out AI in commerce aren't competing on features anymore — they're competing on trust. The 58%, 33% and 71% figures aren't describing a fringe of skeptics; they describe the majority of the market.
Customers aren't psychologically or socially ready to hand purchase decisions to a machine, and economic pressure only reinforces that. In this race, the winner won't be whoever deploys AI fastest — it will be whoever earns customer trust first.
More analysis at digitalforce4retail.com/insights