
When a grocery chain in Germany tested AI-powered product recommendations in 2023, they found a surprising disconnect. While 68% of shoppers admitted they didn’t fully trust the AI’s suggestions, 54% still clicked on them—and 32% ultimately bought the recommended items. The results, documented in McKinsey’s latest consumer behavior report, highlight a critical tension in the AI adoption curve: people use tools they don’t believe in, but only if the trade-off is low effort.
The study tracked 12,000 shoppers across three European markets over six months. Researchers discovered that trust wasn’t binary—it was situational. Shoppers trusted AI more when:
One fashion retailer in the UK took this a step further by letting customers “audit” the AI’s logic. After adding a “Why this recommendation?” toggle, their conversion rate on AI-driven suggestions rose by 19%. The catch? The AI’s explanations had to be under 25 words—any longer, and users disengaged.
This behavior reveals a fundamental truth about AI agents in consumer markets: they don’t need to be perfect—they need to be usable. The McKinsey data suggests that trust is less about the AI’s accuracy and more about the perceived risk of acting on its advice. A 2024 survey by PwC found that 73% of consumers are willing to try AI tools if the downside is minimal (e.g., a free sample or easy cancellation).
For businesses deploying AI agents, the lesson is clear: don’t hide behind “black box” models. Instead, design for transparency at the point of decision-making. The German grocery chain saw a 12% lift in repeat purchases when they added a one-sentence rationale to every AI suggestion. The key was making the AI’s reasoning visible, not necessarily convincing.
What’s next? Expect more brands to adopt “explainable AI light”—simple, digestible justifications that reduce perceived risk without overpromising. The AI agents winning in retail won’t be the most advanced—they’ll be the ones that make users feel like they’re in control, even when they’re not.
Photo: NSYS Group / Unsplash (https://unsplash.com/@nsys_group)
European utilities are increasingly deploying AI to navigate complex energy transitions, focusing on practical applications like grid optimization, demand forecasting, and predictive maintenance to integrate renewables and enhance operational efficiency.

A 200-employee Ohio metal parts factory reduced unplanned downtime by 40% in 18 months using AI agents for predictive maintenance. Here’s how they did it.

Comments (3)
That's fascinating about the 25-word limit for explanations - did the fashion retailer test if shorter explanations led to higher trust or just higher engagement?
That's fascinating about the 25-word limit for explanations. Did the fashion retailer in the UK test if longer explanations would've worked for high-involvement purchases, like electronics?
I'm curious, do you think the 19% increase in conversion rate for the fashion retailer in the UK would have been higher if they allowed users to provide feedback on the AI's explanations, essentially creating a feedback loop to improve the AI's logic over time?