The Stylitics Marketing Team explores the intersection of AI, retail, and shopper experience, sharing strategies and insights that shape the future of product discovery and visual merchandising.
In fashion e-commerce, conversion rates have always hinged on the quality of product images. Product Detail Pages (PDPs) are where shoppers decide to buy—or bounce—and imagery is often the single biggest driver of that choice. Traditionally, brands relied on costly, time-intensive photo shoots with lighting crews, stylists, and models. That system works, but it becomes difficult to scale for global retailers with thousands of SKUs.
AI-generated images are now being tested as a way to expand on-model coverage and reduce production bottlenecks. With the right quality control in place, these tools can help retailers experiment with new types of visuals and explore whether they can improve shopper engagement and conversion.
Key Takeaways:
On-model imagery drives confidence: 76% of shoppers say it is the most helpful format for purchase decisions.
AI imagery serves two shopper needs: inspiration (e.g. versatile “3-styles”) and information (e.g. 360° views).
Quality matters most: 71% of shoppers could not distinguish AI from real photos, but accuracy of details like fabric, fit, and buttons is non-negotiable for trust.
Transparency is key: 59% of shoppers want disclosure when AI is used, interpreting it as a sign of honesty and integrity./
Why PDP Imagery Matters
Product Detail Pages (PDPs) are often the most influential part of the e-commerce journey. They’re where shoppers decide whether a dress drapes the right way, jeans will fit their body type, or sneakers will pair with a jacket already in their cart.
To understand what most shapes those decisions, Stylitics partnered with Aha Studio to survey 411 shoppers across multiple cohorts. The study tested different types of product imagery and styling formats to uncover which visuals build the most confidence and inspire purchase intent.
The findings were clear:
76% of shoppers prefer on-model photos over flat lays or collages because they better communicate drape, proportion, and styling context.
The luxury retailer Milaner tested AI-powered on-model imagery and achieved a 157% increase in conversion rate and a 40% boost in shopper engagement. Results like these show that moving from traditional product photography to AI-powered fashion photography can directly improve shopper trust, engagement, and conversion.
Do Shoppers Trust AI-Generated Images?
Download the full study and gain insight on how 400+ shoppers really feel about AI imagery in fashion ecommerce.
In 2025, Mango became one of the first major fashion retailers to roll out AI-generated product images as the primary visuals on its ecommerce site. Traditional model photography was replaced with AI human models across Product Detail Pages (PDPs), not just in background or lifestyle shots. A disclaimer on each page notes that virtual models are used, but for most shoppers the AI-generated imagery is nearly indistinguishable from real photography. Mango’s demonstrated how quickly AI-powered visual creation can shift from experimental marketing assets to core PDP content.
In contrast, Australian fashion label Atoir faced shopper criticism when it listed apparel on The Iconic using AI fashion models. While the product description included a disclosure—“please note this image is AI-generated”—it was buried at the bottom of the page. Shoppers noticed. Some praised the efficiency, but others flagged concerns over transparency and questioned whether the garments would look the same in real life
Shopper Sentiment: What the Data Actually Says
Mango and Atoir prove that with the increase in efficiency, there is risk. AI-generated PDP imagery may create efficiencies and new visual possibilities, but it can also spark shopper skepticism if accuracy or disclosure falls short.
But how do shoppers actually feel when they see AI images when shopping?
To answer that, Stylitics and Aha Studio surveyed 411 shoppers across web and social channels. The results cut through assumptions and executive fears:
71% couldn’t tell the difference between a real product photo and an AI-generated one.
When told the images were AI, 60% reacted positively or neutrally (“Cool, that’s smart” or “Interesting but not a big deal”).
59% want disclosure, interpreting it as a signal of honesty and brand integrity.
Overall, shoppers are not uniformly rejecting AI imagery. Their acceptance depends on execution quality, transparency, and accuracy, with accuracy of details like fit, fabric, and proportions emerging as the most critical factor for trust.
Two Jobs for PDP Imagery: Inspiration and Information
Our research shows that PDP visuals don’t just sell products—they perform two distinct jobs that influence shopper decisions.
1. Inspiration:“Can I see myself in this?”
This dimension is about sparking imagination and showing how an item might fit into a shopper’s lifestyle. In our research, female shoppers gravitated toward formats that emphasized versatility—like a “3-styles” layout that displayed a dress worn three different ways.
AI-driven tools such as outfit generators and styled layouts are being explored as a way to scale this kind of inspirational content, offering the potential to showcase a single item in multiple contexts more efficiently. For brands, the opportunity lies in experimenting with how new approaches to on-model imagery could expand lifestyle storytelling—from casual to formal, or from day to night—while still maintaining accuracy and brand fidelity.
2. Information: “Will this actually work on my body?”
Here, accuracy becomes non-negotiable. In our study, male shoppers favored rotating 360° views that highlighted details such as stitching, drape, and fabric texture. To build true shopper confidence, any AI-generated on-model imagery would need to deliver highly realistic proportions and fine details—an area where errors risk breaking trust and increasing return rates.
Because the stakes are highest at this stage of the purchase journey, many retailers begin experimenting with AI in lower-risk contexts (such as ad generation or background replacement) before carefully testing on-model PDP visuals. The key challenge is not just generating images, but ensuring the accuracy and quality control required for product-level trust.
The ROI of AI-Generated PDP Images
Retailers are exploring AI-powered product visuals not just for cost savings but to test new ways of driving measurable impact. Early areas of experimentation include:
Speed to market: AI imagery has the potential to shorten campaign timelines, making it easier to test and refresh visuals more quickly.
Basket size: Bundled looks and outfit generators may encourage larger orders, though results depend heavily on execution and shopper trust.
Engagement signals: Richer product visuals and AI-generated backgrounds are being tested to see if they influence how long shoppers spend on-site.
Operational efficiency: By supplementing traditional photoshoots, AI can help reduce some production bottlenecks, though human photography remains essential for accuracy and fit.
A/B Testing: A Careful Way to Explore AI PDP Images
Retailers exploring AI-generated PDP images should take a measured, test-and-learn approach rather than broad deployment. Early experiments often use A/B testing frameworks such as:
Control vs. AI group: Comparing traditional photos against AI-generated versions.
Key metrics: Monitoring conversion, return rates, and CTR. Returns are especially important to track—our research shows 37% of shoppers check return policies more carefully when they know an image is AI.
Channel testing: Some retailers experiment with AI ad or video generation tied to PDP visuals to evaluate performance across marketing touchpoints.
Phased use cases: Tests typically begin with lower-stakes, inspirational imagery (outfits, backgrounds), before carefully moving into core PDP visuals where accuracy and trust are critical.
Risks and Requirements for Success
AI on PDPs comes with both opportunity and pitfalls. Key areas to manage include:
Accuracy risk: Poor AI-generated product images can increase returns. Every detail matters—from patterns and prints to stitching and 3D elements.
Brand consistency: Maintaining brand fidelity and aesthetic alignment is essential, and requires quality control processes tailored to each retailer.
Disclosure: Shoppers expect honesty. A small label such as “AI-generated” or “Virtual Model” can help preserve trust.
Execution quality: While many shoppers are neutral or positive about AI when quality is high, low-quality or misleading visuals risk undermining confidence and driving higher returns.
From Experiment to Expectation
The exploration of AI-generated PDP images is not about replacing human models—it’s about testing whether new approaches can make product imagery more flexible, scalable, and cost-efficient, while still protecting shopper trust and brand standards.
Mango illustrates how some retailers are experimenting with AI visuals in place of traditional on-model photos, while Atoir highlights the risks of poor disclosure and fit accuracy. Shopper surveys reinforce that quality and transparency matter more than whether an image is “real.”
The long-term winners in this space will not be those who simply deploy AI fashion models, but those who integrate AI thoughtfully—combining product photography, AI-powered backgrounds, styled outfits, and ad content into a seamless, trustworthy experience.
AI imagery today is still in the test-and-learn stage, and its broader role will depend on how retailers execute on accuracy, trust, and brand consistency.
FAQ
In our study, 71% said real and AI images looked the same or had only small differences.
60% reacted neutrally or positively, though skepticism centered on fit accuracy and authenticity.
Yes. 59% of shoppers advocated for clear labeling (a small note or icon). Non-disclosure creates trust risk.
Research suggests a phased approach: begin by ensuring a strong on-model baseline, then test AI for inspirational use cases, and eventually explore more advanced applications like fit/personalization—always with strict quality controls in place.
