Key Stats & Takeaways
- 20+ new insights on the influence of AI on shopper habits and preferences
- 2,273 US online shoppers surveyed in this quantitative UX study
- While shoppers are rapidly adopting AI for product discovery, critical hesitation remains around accuracy, imagery, and purchasing authority, making transparent design and easy fact-checking essential for retailers
We’ve released new Quantitative Insights into the perceptions and usage of AI tools in ecommerce, adding to our growing body of data on the habits and preferences of online shoppers.
These insights are survey-based data visualizations that complement and deepen our large-scale UX research findings of shoppers’ AI engagement.
The 20+ insights address the use of AI in the online shopping experience, spanning a breadth of topics: product research entry points, AI feature usage and interest, trust and privacy concerns, perceived accuracy of AI content, confidence in AI-generated images, AI purchasing authority limits, preferences for AI vs human support, and responsibility for AI purchasing errors.
Online shoppers navigate a delicate balance of convenience and trust when encountering AI across their online shopping journey.
While shoppers eagerly adopt AI for fast product discovery and feature comparisons, they remain skeptical when evaluating AI-generated product images or resolving order issues.
Beyond early-stage product research, consumers expect clear boundaries, welcoming AI chatbot assistance for routine tasks while favoring human support the moment an issue arises.
Each of these friction points plays out at a different stage of the journey, and each has distinct implications for how e-commerce sites should implement and communicate AI features.
In this article, we’ll highlight 2 high-level findings that reflect how shoppers navigate these friction points with AI:
- Shoppers prefer AI for low-stakes product research but favor human support for issue resolution
- AI-generated product images reduce purchase confidence for the majority of shoppers
Shoppers Prefer AI for Low-Stakes Research but Favor Human Support for Issue Resolution
While 63% of shoppers prefer AI chatbots for product research tasks like comparing features, roughly two-thirds favor human representatives when resolving order errors. (see Quantitative Insight #AI025 for an interactive version of this graph alongside key findings and actionable insights).
Online shoppers draw a clear line in the sand when deciding where AI fits into their customer journey, showing high willingness to adopt automated assistance for product research while rejecting it for post-purchase problems.
Our latest data reveals that a majority of shoppers prefer AI chatbots for low-stakes information gathering, including comparing product features (63%), checking stock availability (59%), and getting personalized recommendations (56%).
However, when order errors occur, that preference reverses dramatically: roughly two-thirds of respondents prefer a human representative to handle damaged or incorrect items (69%), late or missing orders (67%), and shipping or billing mistakes (66%).
This division of labor underscores that consumers view AI as a convenient research assistant, not a reliable proxy for high-stakes problem resolution.
Forcing shoppers into automated bot loops when an order goes wrong can introduce severe friction, escalating frustration at the exact moment customer retention is most vulnerable.
Retailers should therefore align AI deployment with shopper comfort levels by embedding chatbots natively within discovery and research workflows, while keeping live human support easily accessible.
Ensuring clear, prominent escalation pathways to human agents, or providing transparent self-service options when live support isn’t available, allows sites to leverage AI efficiency without eroding shopper trust when mistakes happen (see Quantitative Insight #AI025).
Note: We also provide qualitative insights into shoppers’ use of AI assistants and how they should be implemented, in our UX guidelines #3458, #3461, #3464, and #3467.
AI-Generated Product Images Reduce Purchase Confidence for Most Shoppers
While 18% of shoppers say AI-generated product images don’t affect them, 60% report that knowing an image is AI-generated reduces their purchase confidence (see Quantitative Insight #AI027 for an interactive version of this graph alongside key findings and actionable insights).
While AI tools offer retailers a cost-effective way to generate visual assets, relying on AI-generated imagery creates a major buying hurdle for online shoppers.
Our June 2026 survey data reveals that 60% of online shoppers say knowing a product image is AI-generated would negatively impact their purchasing confidence.
Specifically, 25% report it would slightly reduce their confidence, 18% say it would majorly reduce it, and 17% state they would not feel confident enough to buy the product at all.
These figures reveal that shoppers view product imagery primarily as a tool for verification and accuracy, rather than mere decorative content.
Retailers should therefore exercise caution when introducing AI-generated imagery, prioritizing authentic photography for core product views.
When AI-generated visuals or lifestyle backgrounds are used, brands must actively take steps to counteract shopper skepticism and safeguard conversion rates (see Quantitative Insight #AI027; check out the top features that restore buyer confidence.
20+ Quantitative Insights to Inform AI-Related UX Improvements
In addition to the 2 insights presented here, all 20+ AI Engagement quantitative study insights are available on Baymard, along with our Quantitative Insights for general ecommerce on Awareness & Product Discovery, Consideration & Purchase Behavior, Checkout & Post-Purchase, Loyalty, and other industry-specific studies.
These are just 2 of the insights from our AI Engagement quantitative study.
Explore the full set of insights to build an in-depth picture of how shoppers interact with AI features and what they expect across the online journey.
Pair these insights with our 700+ UX guidelines based on observed user behavior to ground your UX design decisions in both behavioral research and survey data.
We also offer Quantitative Insights for general ecommerce on Awareness & Product Discovery, Consideration & Purchase Behavior, Checkout & Post-Purchase, and Loyalty along with industry-specific studies into Flight Booking & Airlines, Apparel & Accessories, Furniture & Home Decor, and Home & Hardware.
Getting access: all 20+ AI Engagement quantitative insights are available now within Baymard. (Already have an account? Go directly to the AI Engagement quantitative study).
To find out how your ecommerce desktop site, mobile site, or app performs against industry benchmarks, learn more about a Baymard UX audit of your site or app.
