How AI changes the online shopping experience

Generative artificial intelligence (AI) has become a major player in the online shopping space. Companies are leveraging AI not just to streamline processes but also to revolutionize how products are developed, marketed, and sold. However, the big question remains: Has AI actually improved our shopping experience? To explore this, The Verge's Mia Sato conducted a deep dive into AI-driven shopping platforms and tools, including a try-on session with the AI fashion brand, Finesse, while testing related AI-powered e-commerce tools. Let’s unpack this fascinating journey in detail.

AI's growing role in online shopping

Companies in the e-commerce sector are now deploying AI in almost every stage of their operations. From product creation to customer engagement and sales assistance, AI is believed to be the key to unlocking new efficiencies and experiences.

For instance:

• eBay allows its sellers to automatically generate product descriptions.
• Amazon deploys an "AI fit" feature to assist users in selecting the right clothing size.
• Shopify, a giant in the e-commerce ecosystem, offers AI tools to edit product images and craft website copy.
• Even Levi’s experimented (albeit controversially) with AI-generated models to increase representation and diversity without actually hiring new talent.

According to sellers and platforms, such tools simplify operations and allow for faster, more detailed product listings. Theoretically, AI applications here serve both parties: sellers use fewer resources while presenting polished product pages, and buyers enjoy what’s intended to be a more seamless shopping experience. But how accurate is this hypothesis in practice?

AI's major impact on online shopping tools begins here.

The AI-driven fashion house: Finesse

One of the shining examples, or perhaps experimental prototypes, of using AI in e-commerce is Finesse, an AI-driven fashion brand that utilizes data to identify and manufacture trendy clothing. Established in 2021, Finesse markets itself as the first truly AI-driven fashion brand. Its proprietary algorithms scrape social media platforms and comment sections to pinpoint trends, collect user sentiments, and predict how much inventory to produce, aiming to avoid wasteful overproduction seen in traditional fashion.

Their website experience itself is a mix of the unreal and highly curated. Many of the displayed designs are 3D-rendered, relying heavily on digital fashion concepts rather than actual garment tests. From reimagined versions of celebrity looks, such as Taylor Swift’s Grammy dress, to outfits adopted from viral trends on TikTok, Finesse is effectively recycling the internet’s fashion dreams back into purchase-ready products.

But this system isn’t without hiccups. While Finesse prides itself on predicting buyer needs, out-of-stock items on its platform suggest flaws in its algorithm’s forecasting. For instance, despite claiming precise inventory control, one of the more sought-after designs was unavailable during Mia's session. This raises questions about how effective its AI methods truly are.

Finesse boasts a comprehensive use of AI for modeling trends and generating attire.

Generative AI’s broader impact in retail

Fashion isn’t the only industry where generative AI has left its footprints. Across multiple domains of online retail, AI-generated content is commonplace—sometimes even misleading. On visual-heavy platforms like Instagram, for example, knitting inspiration images are often circulated that are unmistakably AI-generated.

Similarly, on food delivery platforms like DoorDash, many restaurant images reveal dishes that are not real photos but digital renderings. These misrepresentations can confuse consumers about what they are buying, whether it's an unrealistically perfect-looking meal or a craft pattern for something that doesn’t exist in real life.

E-commerce platforms like Etsy and Amazon also face challenges. Sellers frequently post AI-manipulated photos of products like pet sweaters, creating expectations that might not match reality. Yet, these listings still generate sales—proving how easily digital renderings can manipulate consumer trust when they are strategically placed alongside genuine content.

Some AI-generated retail content feels noticeably “off” to trained eyes.

Our Finesse haul: Unboxing and try-on

Here comes the most anticipated part—actually unboxing and trying on clothes from Finesse's AI-influenced collection. According to Mia, while some of the items visually aligned with what was advertised on the website, significant issues quickly began to emerge.

For example:

• Boots: These “convertible” boots looked promising online but felt impractical in real life. Fasteners were located in uncomfortable positions, and the material didn’t give any sense of durability. Mia described them as “how AI would design a boot without understanding human anatomy.”
• Black knit dress: The stitching was visibly pulling apart at seams, and the overall construction screamed “fast fashion quality.” While passable, it offered clear evidence that digital fitting isn’t foolproof when it comes to real-world wearability.
• Taylor Swift dress knock-off: A replica of Taylor’s Grammy look was nearly unwearable due to comfort issues. Its floral accents kept drooping—something easily missed in digital mockups but glaringly obvious in person.

Mia examines whether Finesse’s AI-inspired designs live up to expectations.

This part of the review underscored the risks of skipping essential garment prototyping and in-person modeling.

Selling using AI tools: Pebblely & eBay

Following the unboxing review, Mia decided to offload some Finesse products and explore AI tools designed to enhance resale efforts. Enter Pebblely, a platform that uses AI to create backgrounds, stage mock-ups, and even replace mannequins for product visuals.

While innovative in theory, Pebblely presented its own set of mishaps. Some AI-generated images severely distorted human anatomy (think extra arms), while others didn’t include elements like pants that users explicitly requested. Clearly, AI staging tools remain far from perfect.

Then came eBay’s AI-powered descriptions. Unfortunately, this feature didn’t impress either. The descriptions generated by eBay’s AI leaned heavily on vague, fluffy details about color pairings or general compliments about the item, failing to offer any actual purchasing insight—for both sellers and buyers, misleading descriptions could invite unnecessary friction.

Pebblely struggles to produce realistic AI-generated product displays.

Generated descriptions fail to add meaningful value.

Did AI improve the shopping experience?

After testing Finesse products and AI-powered resale tools, one thing is evident: while generative AI introduces exciting possibilities, there are limitations to its efficiencies in retail environments.

On the positive side:

1. The Finesse items largely resembled their online counterparts. This reflects progress in digital imagery aligning with what customers see in their cart.
2. Tools like Pebblely demonstrate some promise in making e-commerce more accessible to small sellers—albeit with room for significant improvement.

However, the downsides remained glaring:

1. Comfort, quality, and practical usability of AI-designed garments leave much to be desired. Finesse’s reliance on digital prototyping creates a clear gap in ensuring well-made clothing.
2. AI-generated product descriptions—whether on e-commerce giants like eBay or by smaller platforms—read as unhelpful filler content rather than informed selling points.

Ultimately, Mia concludes that AI didn’t enhance her shopping experience—products were speedier to produce and sell, but not necessarily better. As generative AI further infiltrates digital landscapes, conscious consumer awareness will be crucial.

Consumers should stay cautious when evaluating AI-generated or AI-moderated purchases.

Conclusion

Generative AI has undeniably transformed the way businesses and consumers interact within the online shopping ecosystem, offering tools that are simultaneously groundbreaking and problematic. While brands like Finesse prove that fast, AI-driven fashion is possible, their shortcomings emphasize there’s no substitute for real-world prototyping when quality and comfort are required.

For resale platforms and e-commerce tools like Pebblely and eBay, it’s clear that much development is needed before AI tools achieve reliability and value for all users involved.

As we move forward in this AI-shopping landscape, the advice for consumers is simple but critical: Read carefully, choose consciously, and double-check what’s placed in your cart. While AI can make waves, humans remain best-equipped to deliver thoughtful judgment before clicking “Buy Now.”