Amazon Blog Hero Image
Sponsored Content

How Generative AI is Energizing the Beauty Industry

Justin Honaman
Worldwide Head, Retail and Consumer Packaged Goods (CPG) Go-to-Market, AWS
Justin Homaman

The beauty industry, like many others, has undergone unprecedented change in recent years with brands scrambling to keep pace. Before 1990, the consumer experience in this industry was controlled mainly by department stores and grocery and drug outlets. Big brands like Chanel, L’Oréal, and Revlon dominated the market. 

From 1990 to 2015, brands like Sephora, Ulta, and Target flourished. During this time, self-shopping and “masstige” (prestige for the masses) blurred the lines between premium and mass market brands. There was also a rise in “indie” and “clean” brands, such as Benefit, Bliss, Tarte, bareMinerals, AXE, Honest Company, and Method. 

Then starting around 2015, the industry began integrating new digital strategies and platforms. Technology started to disrupt the legacy industry with the launch of many direct-to-consumer, digital-native brands (DNBs) and the advent of new social shopping platforms like Instagram and TikTok. 

Marketplaces like added new storefronts to differentiate brands while Fenty and celebrity-influencer brands raised the bar on inclusivity. By 2026 it is projected that nearly 30% of beauty sales will be online. 

This is a dramatic channel shift.

New Technology Powering Beauty

Beauty companies in 2023 are focused on a host of issues: 

  • Meeting volatile demand in multiple global markets—resulting in challenges to forecasting and replenishment
  • Returning to brand and product innovation—emphasizing inclusivity, clean, sustainability, and being an influencer
  • Meeting the customer where they shop—including “phygital” (physical plus digital) and social platforms 
  • Embracing new marketing—influencers can be any customer with a voice
  • Environmental, social and governance (ESG)—consumers expect transparency and ethical, environmentally conscious, cruelty free products

The Amazon Premium Beauty Store, which launched in 2013, and more recently the beauty component of the Luxury Store, which launched in 2020, includes curated luxury brands such as Dr. Barbara Sturm and Clé de Peau Beauté. With retail brands and their customers at the center of its Premium Beauty strategy, Amazon uses innovative digital tools and platforms to elevate the shopping experience. For example, Amazon customers rely on the shopping platform to conveniently search for, compare, virtually try-on, and match complementary products. Millions of beauty customers use as a starting point and search engine to learn more about beauty brands. 

Amazon set up Amazon Beauty as a trusted shopping destination where customers can be inspired, discover new brands, and browse a vast selection of products. The storefront encompasses thousands of brands, and that number is growing with new additions including Lancôme and Shiseido. Each has their own unique brand store within Amazon Beauty. 

Technology plays a major part in creating an immersive, interactive, and personalized Amazon Beauty shopping experience. Machine learning (ML) technology powers the platform’s ecommerce search and site personalization capabilities, delivering the right customer experience at the right time. Shoppers can evaluate beauty products on Amazon’s livestreaming platform—Amazon Live—which combines entertainment with the shopping experience. Amazon Live allows shoppers to chat with other like-minded shoppers, ask the host questions, and purchase deals in near real-time (or while watching the playback later).

Going forward, shoppers can expect to see innovations such as: 

  • Augmented reality (AR)-driven virtual try-on
  • Personalized skin assessments
  • More livestreaming (already ubiquitous in Asia-Pacific markets)
  • ML-based shade matches and recommendations
  • Live and AR virtual service assistance
  • Generative artificial intelligence (AI)-powered stylists
  • Matched product recommendations 

Generative AI in Beauty

Let’s dive deeper into the use cases for generative AI in the beauty industry. Generative AI is a type of AI that can create new content and ideas, including conversations, stories, images, videos, and music. This new content can range from creativity to productivity and business insights. Like all forms of AI, generative AI is powered by ML models—very large models that are pre-trained on vast amounts of data are commonly referred to as foundation models (FMs). 

It's important to note that generative AI leverages the latest advances in ML. It’s not magic but represents the latest iteration of a technology that has been evolving rapidly. 

What makes FMs special is that they can perform many more tasks because they contain a larger number of parameters capable of learning complex concepts. Through their pre-training exposure to internet-scale data in all its various forms and patterns, FMs learn to apply their knowledge in a wide range of contexts.

It's still early in the generative AI craze and the use cases in beauty will continue to grow and evolve. As ML adoption spreads, we are going to see more opportunities for generative AI to reinvent every customer experience and application. We expect personalization—a big selling point for beauty and skincare brands—to gather momentum as businesses deploy generative AI models to increase creativity and productivity. 

Many AWS beauty customers are now evaluating how to leverage generative AI across a broad range of use cases for advertising and marketing. These include:

  • Driving creative and content development
  • Enhancing customer journeys
  • Personalizing advertising and marketing experiences
  • Understanding customer sentiment
  • Optimizing business operations

A good use case involves creative and media planning. With the explosion of digital content and media touchpoints, it’s more important than ever for personalization and optimization to happen as close to real-time as possible. 

Increasingly agencies are focused on helping beauty companies customize the design of different creative permutations with optimized KPIs for different media. They are using generative AI to help optimize the work of design teams by quickly adapting the sizing, titles, and formats to all different networks and platforms. 

This enables teams to produce dynamic creative assets that respond in real-time to consumer trends, driving greater engagement and ROI.

We know that personalized product and content recommendations tailored to an individual beauty customer’s interests are more likely to drive conversion. Yet, we haven’t seen virtual fitting rooms take off at scale. Generative AI can make it easier for consumers to generate their unique body measurements to see how an item will fit on their bodies. This will be a real game changer for the customer experience, while reducing the number of returns to ecommerce retailers who stock a variety of brands with multiple size charts. 

There is also a real opportunity to better assist customers looking for products in the store. For example, generative AI can be integrated with web and mobile apps to help in-store employees more accurately respond to requests like “What kind of dress do I need for a wedding in Southern California?” Employees could also query category and product-level information to provide recommendations for products located outside their department.

Generative AI is not going to completely replace the imagination of beauty designers and marketers, but it could help them overcome a creative block and accelerate time to market. 

Today it’s possible for product innovation teams to input product details, ingredients, brand designs, historical product lines, and forecasted trends into a platform powered by generative AI. The platform would automatically create an array of options, allowing beauty retailers to experiment with an enormous variety of ideas for new collections.

Think about the process improvements that can be built into the design process for beauty products. For example, several of our customers in the industry are close to using augmented reality to visualize beauty products on a variety of body types and sizes. There is also immense potential for using generative AI to reduce the time and resources needed to design products while cutting back on materials used in the manufacturing process. Meanwhile, sophisticated brand managers can use generative AI to query complex datasets using natural language to spot fashion trends and identify innovative new products.

From Ideas to Action

Generative AI has arrived and the beauty industry will never be the same. Try using GENAI to move from ideas to action.

G.  Get going. Move now to assign resources to at least understand generative AI’s potential and how you can benefit. 

E.  Engage. Test, try, learn, adjust, commercialize. Generative AI will power innovation acceleration. 

N.  Net it out in business and tech terms for your staff. This is not an "IT thing." Or an "IT project." It's a business enabler and transformational capability. 

A.  Ask questions. Ask to understand. Ask to dive deeper. Ask to improve, refine, and to make better ML models. 

I.  Make "Intelligence" a personal obsession. Build your own intelligence around AI, ML, generative AI, and more. You'll be glad you did.


More Blog Posts in This Series


    How Generative AI is Energizing the Beauty Industry

    Learn how beauty companies are evaluating the best use of generative AI across a broad range of use cases for advertising and marketing.
    Amazon Blog Teaser Image

    5 Tenets of a Successful Data Products Strategy

    Integrated data and analytics have been a top investment priority for many years—a trend that has only been accelerated by flexible, agile, scalable cloud platforms and new analytic capabilities that did not exist in the on-premise environment. Here are five keys to building and maintaining a successful Data Products strategy.
    Hey Ankit, the test will need to be sent from the preview portion of Marketo From there you can change the 'default' to person You'll have to enter in everyone's email one by one and send a test that way