3 ways fashion ecommerce will evolve through generative AI and ChatGPT

While retail as a whole is looking for ways to deploy generative AI to support operations and enhance customer experience, fashion ecommerce is likely to see some of the biggest impacts. Here’s why:

• Major fashion brands are early adopters of new tech. They are among the first (and most active) to experiment with emerging concepts and tools such as Web3, the metaverse, AR, and NFTs. ChatGPT and generative AI are no exception.
• Fashion retail is heavily dependent on ecommerce. Apparel and accessories are the largest single category for US ecommerce, currently making up 18.7% of all digital sales (worth nearly $215 billion in 2023). Around 36% of US apparel and accessories sales will take place online this year, per our forecast, and that share will surpass 46% in 2027.
• Fashion brands need to solve their returns problem. The challenges of buying clothes online—mainly fit and sizing issues—resulted in nearly a quarter of US apparel and accessories retail sales being returned in 2022, per our forecast. Consumers increasingly engage in bracketing (ordering multiple sizes and returning whatever doesn’t fit) and the costs associated with this practice are unsustainable for many retailers.

McKinsey estimates that generative AI could add between $150 billion and $275 billion to the fashion industry’s operating profits over the next three to five years.

Here are three key ways that brands can use generative AI in a customer-facing capacity.

1. Product search and discovery

Fashion brands already use machine learning to analyze consumer data and provide product recommendations, but generative AI has the potential to create hyper-personalized experiences, leveraging direct and immediate feedback to improve search and discovery.

• Generative AI tools can help shoppers find products through personalized recommendations and conversational search functions. The interactive nature of generative AI exchanges will lead to more engaging experiences. According to Capterra’s February 2023 “Retail Chatbots Survey,” 67% of ChatGPT users said they “often” or “always” felt understood by the bot, compared to 25% of retail chatbot users.
• As generative AI tools learn more about individual shoppers, recommendations will become more relevant, and interactions may become more like those with a stylist or personal shopper. Klarna recently partnered with OpenAI to launch a ChatGPT plugin that offers curated product recommendations and links to purchase through Klarna’s search and price comparison tools.
• Generative AI for product search has the potential to replace existing search filters by allowing fashion consumers to make specific queries about exactly what they are looking for using natural language.

Risk potential: Brands need to have enough high-quality data to support relevant recommendations, and to stay on top of maintaining it. Brands need to ensure that proprietary data isn’t shared in a manner that could compromise its security.

2. Product description and information

Generative AI tools that produce images from text and vice versa will enhance product description pages with richer and more personalized content, highlighting information that is most relevant based on specific consumer needs and preferences. Generative AI for video is advancing rapidly and will allow better 360-degree content to be created from still images.

• Shopify’s OpenAI-powered tool generates product descriptions with just a few details, such as choice of “tone” and SEO keywords, and it plans to integrate generative AI into more features.
• Stitch Fix has trained its generative AI model by having its staff write hundreds of product descriptions based on specific prompts, and the company reported that the resulting descriptions produced by AI obtained higher-than-average quality scores.

Risk potential: The collection of personal data and role of virtual models can be controversial. Levi’s recent announcement of plans to use AI-generated models drew criticism for failing to advance real diversity. Transparency and quality control are critical when using generative AI tools to create content. More than 70% of respondents in a March 2023 dentsu survey agreed that brands should disclose the use of AI in customer-facing functions. All assets should be reviewed for accuracy and bias.

3. Product customization and co-creation

As more people explore the creative potential of generative AI tools, brands can engage them by offering opportunities to co-create products.

• During Metaverse Fashion Week last month, Tommy Hilfiger hosted an AI design contest to create a digital fashion item in the brand’s style, with the winning design produced into a digital collectible by virtual fashion platform DressX.
• The upcoming AI Fashion Week in New York, backed by online retailer Revolve, is drawing a mix of professional and amateur designers to experiment with creation tools like Midjourney to conceptualize styles that can be produced in real life.
• Generative AI tools will allow shoppers to easily customize or suggest products they would like to see. A new wave of AI startups will enable micro-brands to flourish, and an increase in made-to-order and low-volume production can help reduce waste, inventory risk, and returns.

Risk potential: Fashion brands may not be ready to share control over design with consumers, but creators are already using generative AI tools to share imaginary branded goods. Brands will need to tackle how to address intellectual property issues associated with generative AI.

This was originally featured in the Retail Daily newsletter. For more retail insights, statistics, and trends, subscribe here.