AI is moving omnichannel closer to the customer

By Eva Ponce and Ken Cottrill

Omnichannel supply chains personalize online retailing by meeting consumer demand for multi-channel shopping. However, the level of customization must increase substantially with the emergence of AI-driven innovations such as product recommendation services tailored to individual buyers.

As these applications grow in both scope and sophistication, e-commerce will become even more personalized, putting more pressure on retailers to develop supporting omnichannel supply chains.  

Preference for personal

Each fall, the Omnichannel Supply Chain Lab at the MIT Center for Transportation & Logistics conducts its Annual Omnichannel Survey (the 2024 survey was launched in October, see the link below). More than 130 logistics, supply chain, and warehousing professionals took part in the 2023 survey. One question asked respondents where they see AI having the most impact on omnichannel fulfillment. The top-ranked areas in terms of AI having a high impact were demand forecasting (79% of respondents), personalization (73%), customer service and chatbots (70%), and inventory management (69%).

These findings reflect retailers’ efforts to keep pace with AI’s impact on online customer expectations. Over the past year, these efforts have intensified, and we expect this trend to continue in 2025.

Such demands are especially evident in younger age groups. For example, in the first quarter of 2023, research company Statista surveyed more than 17,000 consumers in multiple countries about their buying preferences in 2023. The generational cohort that desires personalized product recommendations the most when making online purchases is Gen Z (47%), closely followed by millennial shoppers (46%). Four out of 10 Gen X consumers also value online personalized product recommendations.

Deloitte’s 2024 Customer Loyalty Survey of some 9,800 consumers in different countries concludes that personalization is an opportunity for improvement in the online shopping space. Only 60% of the surveyed consumers expressed satisfaction with the customized and targeted offerings offered by retailers.

Pandering to customer needs

Retailers are responding with a wave of innovation, particularly at the point of sale. Here are three examples.

Unilever’s advice platforms. Unilever has introduced BeautyHub PRO, an AI-powered selfie tool that offers advice to buyers of its skincare and haircare products. Consumers complete a quiz and share a selfie. The platform uses AI to “see” and assess up to 30 visual data points, and make personalized product recommendations based on several Unilever brands.

The company claims that the total shopping basket value of consumers who purchase from BeautiHub PRO is 39% higher than that of customers who shop elsewhere. Also, platform users who discover new recommendations are 43% more likely to complete a purchase, versus consumers who browse on other channels. Unilever says more than 80,000 individuals tried BeautyHub PRO in 2023, and it reached more than 3 million people.

In 2024, Unilever also launched a diagnostic tool for consumers called Scalp + Hair Therapist under its well-known Dove brand name. Customers can access this AI-powered interactive tool via the Dove website. It generates scalp and hair profiles for customers and recommends relevant products. The service includes links to buy the recommended products on Amazon, providing shoppers with a seamless journey from their first interaction to the point of purchase.

Walmart time saver. The average American family spends nearly six hours per week on household planning and shopping, says Walmart. The company’s generative AI product recommendation tool, GenAI Search, can give them back this time by providing product recommendations tailored to specific buying needs it claims. Customers key in descriptions like “What supplies do I need for a newborn?” and the tool creates a list of items. GenAI also helps the retailer better understand customer intentions based on such queries. The company is testing ways to condense product reviews into concise summaries that help customers match products with the features they are looking for. 

The rise of “click & try” options. Fashion retailer Zara offers customers digital fitting rooms where they can try on items virtually before making buying decisions. The company is also using AI technology to enhance the in-store buying experience. For example, hi-tech mirrors in fitting rooms can scan apparel items and recommend matching products from the store.

Another example of this type of technology is Walmart’s View in Your Home AI and machine learning service, which allows customers to virtually place home goods items in their living spaces before making purchases, enhancing the buying experience. Home goods company Ikea offers a similar tool. These tools are meant to work in conjunction with AI chatbots to offer customers their own design assistant.

Supply chain responses

To support innovations like these, retailers are leveraging AI in combination with other technologies to enable their omnichannel supply chains to deliver on the popularity of customized purchasing.

For example, Swedish retailer H&M captures information from search engines and blogs as part of a broader, AI-enabled strategy to improve demand forecasting. Using the technology to refine forecasting models and improve retailers’ ability to anticipate changes such as seasonal demand variations is key to meeting customer preferences for customization.

Companies are also using AI to improve inventory management, for example, by strategically positioning inventory using algorithms. A tool developed by one retailer analyzes demand trends and helps the company set prices dynamically while considering market shifts.

A key area of development is product replenishment. Online buyers are notoriously intolerant of stockouts, but more so if sellers provide them with tailor-made lists of recommended products that they then discover are not available. AI technology is taking replenishment to another level.

Consider, for example, technology that Unilever is introducing in its ice cream business. The company has a fleet of some three million ice cream freezer cabinets in shops, airports, and supermarkets around the world. It is equipping these freezers with AI-enabled replenishment systems as part of a strategy to ensure that best-selling products are always in stock.

The system captures photos of a freezer’s contents and sends them to the cloud, where an AI trained to identify the company’s ice cream brands and determine the current stock level analyzes them. An algorithm generates an order for the distributor if required.

Unilever is working to add the technology to 30% of its top-selling cabinets for priority markets. It aims to equip 350,000 freezers in the next two years. The company says its innovative replenishment system has achieved incremental sales growth in various countries, including the U.S. (12%) and Denmark (30%). Also, the innovation yields more detailed information on buying patterns that provide more upselling opportunities.

Walmart is also looking to extend its inventory management reach into customer domains. The retailer is studying the purchasing patterns of select customers to replenish their refrigerators with essentials automatically when needed. The InHome Replenishment service places orders when items are in short supply.

Challenges and opportunities

Adapting omnichannel to a highly-customized buying experience comes with various challenges. These include integrating disparate information systems, generating the data needed by these data-hungry systems, and recruiting talent with the requisite supply chain and tech skills. 

Retailers also need to work with partners as the race to personalize intensifies. Unilever says it has “developed an advanced AI-driven customer connectivity model that is helping to deliver a truly integrated end-to-end supply chain” in collaboration with trading partners. The model integrates forecast and actual sales data. The company maintains that “removing the traditional barriers between supply chains and increasing visibility of data, the model creates a seamless ecosystem, joining one supply chain to another.” 

* To see MIT Omnichannel Supply Chain Lab’s 2024 Omnichannel Survey visit: mit.co1.qualtrics.com/jfe/form/SV_0eqzo1yRGQTNqHI?Source=SCMR

About the authors

Dr. Eva Ponce is director of the MIT Omnichannel Supply Chain Lab. She can be contacted at [email protected]. Ken Cottrill is editorial director at the MIT Center for Transportation & Logistics. He can be contacted at [email protected].

SC
MR