FullStack Labs

Please Upgrade Your Browser.

Unfortunately, Internet Explorer is an outdated browser and we do not currently support it. To have the best browsing experience, please upgrade to Microsoft Edge, Google Chrome or Safari.
Upgrade
Welcome to FullStack Labs. We use cookies to enable better features on our website. Cookies help us tailor content to your interests and locations and provide many other benefits of the site. For more information, please see our Cookies Policy and Privacy Policy.

From Recommendations to Immersive Try-Ons: AI's Role in Consumer Goods

Written by 
Juan Nino
,
AI Principal
From Recommendations to Immersive Try-Ons: AI's Role in Consumer Goods
blog post background
Recent Posts
Mastering Agile Project Management: Navigate Success with a Large Team
Driving Excellence in QA: Continuous Improvement and Postman Best Practices
Goroutines in Golang for High-Performance Concurrency

In this article, we are pulling back the curtain to reveal the inner workings of AI's role in the shopping landscape. From unobtrusive product recommendations to the intricate tech that powers it all, we will explore how AI is reshaping the consumer goods world.

Table of contents

If you are a tech-savvy reader or have ever bought on Amazon, you probably have heard about the role of AI in recommending products. The ever-growing use of customer profiling and recommendation engines has shifted the industry worldwide. We now live in the so-called age of “hyper-personalized experiences” when spending our money on treats and needs.

However, key market players' recent adoption of generative AI and other flavors of artificial intelligence has upped the game. The “old” days of AI quietly operating in the background of an e-commerce website are gone, and the new reality for companies in the Consumer Packaged Goods (CPGs) industry involves ensuring users and consumers are fully aware of their AI advancements.

Understanding AI in the Consumer Goods industry

The primary purpose of AI is to automate tasks that typically require human intelligence, accomplishing them more quickly and on a broader scale. This is how it benefits all industries. In the Consumer Goods sector context, AI has been like that good old friend who has had a great experience with a product and picks up the phone to get us excited about it and motivate our purchase. This use of AI is known in the machine learning field by the name of recommendation engines.

A recommendation engine is an algorithm that, much like a close friend, uses past experiences from users with specific items to predict your likelihood of favoring similar choices. Based on your affinity, it then customizes the product to match your needs through direct marketing strategies and motivates you to purchase the item. In other words, a recommendation engine uses information from you and people with similar interests to suggest products and promote cross-selling and up-selling.

Comparison of AI recommendation engines: (a) User similarity-based engine. (b) Product similarity-based engine
AI customizes suggestions by pairing you with like-minded users while also broadening your options based on your preferences. Source: Ni et al. (2021)

As I mentioned earlier, this type of AI and data-driven insights have played a crucial part in the world of CPGs over the past decades. The often quoted and humorous “Beer and Diapers” urban legend from the early '90s, which has never been confirmed, illustrates the potential of data analysis. This legend suggests a correlation between increased beer sales and diapers on Thursday evenings, speculating that fathers buying diapers on their way home from work also purchased beer for the weekend. However, it's important to note that this story is usually used only for illustrative purposes. Modern recommendation engines did not exist back then, but they have since been embraced by companies like Netflix and Amazon and have had a significant influence on the CPG industry.

While there are several other use cases where major companies in the industry have harnessed AI (such as optimizing supply chains, embracing digital transformation, and customizing user experiences), the spotlight has primarily been on recommendation engines. This remained the case until very recently.

A new frontier in AI

Over the past decade, all the ingredients for the perfect storm have come together to boost the development of robust AI applications. On one hand, computer power has exponentially grown, and we are on the verge of adopting quantum computing beyond scientific research. Additionally, the volume and availability of data—thanks to the widespread dissemination of the internet and connectivity—have provided an unprecedented knowledge base for AI models to consume and actively learn from. Finally, academia and industry have heavily invested in research around the mathematical principles supporting AI, challenging each other to create ever-improving artificial intelligence systems.

If you combine these elements, you get an accelerated industry actively pursuing to achieve generalized artificial intelligence. As a result, we have seen over the past few months the massive release of a handful of AI tools that are more clever than we expected. From ChatGPT to the LLaMA family of models released by Meta Research, every person with internet access can now use a general-knowledge Q&A bot. On the other hand, with models such as Midjourney, DALL-E or Stable Diffusion, people can bring the craziest pictures to life only by describing what they would like to see.

Having all this technology available and within reach, corporations have not hesitated for a second to adopt and adapt it to meet their needs. We have seen a significant number of companies replacing their Customer Support teams, either entirely or partially, with AI. Business disruptors are piloting conversational AI to support and boost sales and moving from a human-touch-based experience towards an exciting AI-human interaction. Even the film industry is plotting to replace voice actors and extras with AI-generated content, which has both been groundbreaking and controversial.

With the same goals in mind, the consumer industry has kept up. The usual approach of implementing AI in the background has heavily shifted, and artificial intelligence is now delivered directly and openly to consumers. Let’s find out how.

AI: the perfect fit for virtual try-ons

One of AI's most captivating applications in consumer goods, particularly in the fashion industry, is using artificial image generation (AIG) for home-based product experimentation. In essence, AI empowers manufacturers to enhance the customer experience without drastically changing their value chain while reducing the cost of delivering demo products. Trying on shoes or shirts using just your phone camera and a mirror is already impressive. Still, beauty brands are pushing the envelope further to offer an even more thrilling experience.

Makeup brands commonly face the challenge that models in magazines, colorful palettes, and demo products lack the personalized touch that clients expect when trying to find the product that best fits their needs. Early virtual try-ons used basic 3D models and simple interfaces. They allowed users to take selfies and try different makeup, but these experiences felt artificial and lacked realism. Clients could take a selfie with their smartphone and test different lip glosses, shadows, and other beauty care products to see how they looked. However, that technology had its share of limitations; it often struggled to deliver a convincing result. In some instances, users felt as though they were merely playing with a costume makeup app rather than engaging with genuine products—a somewhat underwhelming experience.

With the introduction of artificial image generation into makeup and skincare applications, customers now have the opportunity to explore and experiment with a variety of products in an environment that more closely resembles the real world. AIG brings hyper-reality and immersion to the forefront, ensuring users are presented with virtual try-on experiences that are as close to the real thing as possible. Clients can confidently explore and experiment with an array of products while enjoying an unparalleled level of personalization.

Image showcasing Maybelline's AI-powered virtual try-on technology for lifelike makeup trials.
Maybelline, along with other brands, offers a lifelike makeup trial experience with its AI-driven virtual try-on technology. Source: Harper’s Bazaar.

This exciting AI technology is typically created using “before and after” pictures from several different clients using their products. The model is then trained with the images and features of each product in the company’s portfolio to create a personalized generative AI tool. Such AI models help companies reach more customers, turning each client into a model and providing them with a unique experience. Shopping isn't just about buying; it's about enjoying the journey, and generative AI adds a sprinkle of charm to every step, making each interaction as delightful as it is individualized.

Before-and-after comparison of a woman's skin, illustrating the effects of skincare products.
In 2023, Haut.ai introduced SkinGPT, which uses generative AI to simulate skin changes from skincare product use. Source: Haut.ai 

AI's expanding horizons: challenges and opportunities ahead

As you probably know, there are several other examples of the application of AI in the CPG industry. In today’s landscape, chatbots are managing customer interactions, products are being fine-tuned to meet individual customer requirements, and AI-enhanced goods are reaching consumers directly. This progress paints a complex picture for business leaders. On the one hand, it's an exciting era filled with opportunities; on the other, it's an overwhelming context for those unsure of how to engage effectively with AI technologies. This duality naturally leads to two pressing questions: What's the next frontier for AI, and what challenges must we prepare for in this rapidly evolving ecosystem that spans multiple industries?

While it's tempting to say "the limit is your imagination," the reality is more complex. Rather than proactively embracing innovation, many companies find themselves reacting to the surging AI wave. Unthinkingly implementing AI can be more detrimental than not using AI at all. The key to harnessing the full potential of AI lies in implementing a well-thought-out strategy that effectively bridges your core business model with the enabling technology. 

Here are essential steps to consider:

  1. Develop a Vision. Initiating an AI project without a cohesive strategy can be costly and counterproductive. Clearly define your vision, pinpointing your organization's current position and what AI aims to achieve.
  2. Prioritize use cases. Identify and prioritize use cases that align with your business objectives, creating a portfolio of AI capabilities.
  3. Focus on implementation. Shift your attention towards this phase, leveraging the right tools and planning frameworks like CRISP and AI Project Canvas. Remember, technology alone won't guarantee success—it's the alignment with your business objectives that truly matters.
  4. Seek guidance. At this pivotal stage, professional guidance can prove invaluable. FullStack Labs offers not only top-tier developers and project execution but also expert advice, ensuring the seamless integration of AI into your business landscape.

As you steer your organization into the exciting realm of AI, keep in mind that strategic planning and expert guidance are your compass. By meticulously aligning your AI initiatives with measurable business outcomes, you will not only navigate around common pitfalls but also chart a course toward resounding AI-driven success.

Juan Nino
Written by
Juan Nino
Juan Nino

Juan is an AI engineer with a Master's degree in Industrial Engineering and 8 years of experience leading AI and machine learning initiatives across diverse markets in Latin America, Europe, and the United States. Specializing in AI-based solutions, Juan is proficient in open-source programming languages and cloud ecosystems like Google Cloud, AWS, and Azure. His work revolves around advanced machine learning models, particularly in Natural Language Processing, using state-of-the-art algorithms such as GPT, BARD, and BERT. In addition to his professional roles, Juan has contributed to academia through research and as a graduate assistant.

People having a meeting on a glass room.
Join Our Team
We are looking for developers committed to writing the best code and deploying flawless apps in a small team setting.
view careers
Desktop screens shown as slices from a top angle.
Case Studies
It's not only about results, it's also about how we helped our clients get there and achieve their goals.
view case studies
Phone with an app screen on it.
Our Playbook
Our step-by-step process for designing, developing, and maintaining exceptional custom software solutions.
VIEW OUR playbook
FullStack Labs Icon

Let's Talk!

We’d love to learn more about your project.
Engagements start at $75,000.

company name
name
email
phone
Type of project
How did you hear about us?
Thank you! Your submission has been received!
Oops! Something went wrong while submitting the form.