The strategic power of Personalized AI in the shopping experience

Personalization is not an option, it is an expectation. According to data from McKinsey, 76% of consumers seek personalized experiences and are more likely to buy from brands that offer them.

In addition, eCommerce businesses that adopt personalization strategies can increase their revenue by 10% to 15%. This was one of the main ideas shared by the Amazon Web Services (AWS) team during their presentation at LogiCommerce Connect 2025.

Personalization is in Amazon’s DNA

Amazon began personalizing customer experiences in 1998 by recommending books based on other users' behavior. Today, this capability has evolved to cover all user touchpoints—from personalized recommendations on Prime Video to fully automated sales systems powered by artificial intelligence.

“The goal is clear: improve the customer experience and, in doing so, increase revenue,” the speakers emphasized.

Every touchpoint can be personalized

During the presentation, AWS highlighted that every section of an eCommerce site can be personalized: titles, images, product descriptions, banners, and more. You can even show different versions of the same product depending on the customer profile—for example, displaying a jacket in a snowy setting if the user is searching for mountain getaways.

And if the eCommerce is integrated with physical stores, it's even possible to send automated messages when the customer enters the Geofence area of a store, offering personalized promotions or reminders based on their purchase history.

Four key areas where personalized AI makes an impact

AWS structured the benefits of personalization into four main pillars:

• Acquisition and retention
• Improved engagement
• Operational efficiency
• New user acquisition

They also presented a maturity model for personalization, based on five levels per pillar—from basic customization to generative AI modifying images and content dynamically.

Real Case: Pomelo, fashion leader in Southeast Asia

One of the most compelling examples shared was Pomelo, a fashion platform in Southeast Asia that has implemented Amazon Personalize.

Thanks to this infrastructure, they moved from a general country-level recommendation system to real-time, one-to-one recommendations based on each user’s actions. The results speak for themselves:

• 60% of product clicks now come from personalized areas of the eCommerce site
• They've seen a 10% to 15% increase in revenue from these improvements

Live Demo: From visual AI to textual personalization

In a live demo, Guillermo Bravo (Solutions Architect at AWS) showcased highly visual examples:

• Image recognition to redesign spaces (such as a living room) in styles like minimalist, rustic, or Scandinavian—automatically generating product recommendations from the catalog to replace or complement existing decor.
• Personalized product descriptions based on user profiles. For instance, the same plant is described differently for María Smith (21 years old, landscaping enthusiast) than for Michael (38 years old, interested in interior design), using prompts that adapt content to the detected user persona.
• Product recommendations tailored to customer interests and purchase history, completely modifying the order and types of products displayed.

Application in physical environments

The presentation also explored the potential of geolocated personalization: with the user’s consent, it’s possible to trigger personalized messages when the customer enters a store’s Geofence area—sending relevant offers via SMS, WhatsApp, or email.

What about large catalogs?

When asked about the scalability of these solutions for stores with extensive product catalogs, AWS gave a clear answer: there are already ready-to-use functional blocks, like those provided by LogiCommerce, that enable efficient personalization—even without the immediate need for AI.