Retail executives from leading merchants and specialist technology providers returned to Paris for the second European event from the US-based National Retail Federation (NRF) that showcases cutting-edge technologies and provides a platform for the latest industry thinking.
Not surprisingly at such events, artificial intelligence (AI) is inescapable, but what was also evident this year at NRF Europe was the growing body of evidence that larger organisations are integrating AI-powered solutions alongside undertaking wholesale changes to their underlying technology infrastructures.
This has certainly been the case at Debenhams, which offloaded its physical stores and has adopted an online marketplace model selling products from 25,000 different brands and retailers, which does not involve holding any products in warehouses.
Dan Finley, CEO of Debenhams, says: “It’s been a fundamental transformation of the business model and it’s worked really well. We now see ourselves as a tech-led business. We’ve completely changed all the technology over the lpst two years. It’s delivered greater efficiencies.”
On top of the core IT systems, AI is being used to revolutionise areas such as the merchandising function, with price changes now possible across the 25,000-30,000 brands’ products in only minutes, and website content including descriptions and images can be created using these tools rather than involving a whole team of people.
Debenhams has also put a virtual assistant into the hands of all its customers that enables them to navigate through the company’s now massive inventory and focus in on exactly what they are after. “We are seeing much more expansive search terms and questions being asked. As agentic commerce company’s monetise their services [through advertising], then it will be an evolution for customers,” says Finley.
Home décor and furniture retailer Westwing has also undergone a fundamental technology overhaul. Andreas Hoerning, CEO of Westwing Group, says: “We shifted from our own tech stack to Shopify … moving from our own stack to a software-as-a-service [SaaS] model was a tough decision. As this model has evolved, there have been more off-the-shelf solutions available – whereas 15 years ago, you had to build things yourself. While the front end is still handled in-house all the other technology [for the back end] is SaaS, which simplifies things and enables us to stay ahead with our technology.”
This change has helped it shift its business model away from selling third-party brands to being a retailer of mainly own-label goods. He says 65% of its sales are now generated from in-house designed own-label products (accounting for 90% of margins) across around 5,000 lines compared with the previous 100% of goods from third-party brands, with their lesser margins.
A similar journey is being undertaken by Polish-based Zabka Group that has both repositioned itself as a more modern, dynamic convenience store – attracting a younger audience to its ready meals and hot food-to-go ranges – and a transformation of its technology that has driven great efficiencies and given it the flexibility to build capabilities such as a “super app”.
Tomasz Blicharski, incoming CEO of Zabka Group, says this move has helped it attract top technology talent who have constructed a solution for running the business much more efficiently. At the heart of this is the creation of data lakes, on top of which services have been built such as pricing, assortment and transportation.
“We’ve moved to higher margin products and differentiated formats with different ranges. Every single store now has its own unique planogram that’s created automatically based on the data,” he says, adding that differentiated pricing is also used across its stores that now number 13,000 in Poland and 300 in Romania.
The swing towards electronic pricing labels
Pricing has certainly risen up the agenda of retailers and the power to be able to use data to determine prices more intelligently and change them swiftly is now regarded as extremely attractive. Hence all the major grocery chains around the world are rolling out Electronic Shelf-edge Labels (ESLs) along with the back-end pricing intelligence tools to better use price as a lever to improve efficiency and boost margins.
Like other businesses, Morrisons has found ESLs can play an increasingly important role beyond simply changing prices quickly. Gordon Macpherson, group productivity director at Morrisons, says: “The original premise was to save time, because we have 5,000-6,000 prices changes made on an evening … but as we continue to see online sales grow, then we will see how we can use the tech for quick-commerce. We get the benefits when all the tech interplays in-store.”
What attracted attention at NRF Europe was its use of the labels for “pick-by-light” that has been piloted to help colleagues navigate around its stores when picking for online orders. They are assigned one of seven colours and the label flashes this colour on the relevant product that they need to pick next for the order or the quick-commerce deliveries they are fulfilling.
The pilot has increased pick rate efficiency by 10% – with improvements in both speed and accuracy – that has given Morrisons the confidence to roll it out and give customers the capability to find products on the shelf via the flashing labels. Macpherson has noted the success by Walmart in the US, and so Morrisons will soon be enabling shoppers to select a product within the supermarket’s app and for this to prompt the label to flash.
Within the Innovators Showcase area of NRF Europe, AI solutions were certainly prevalent and in various instances sought to bring together disparate data sources and inject large language model (LLM) and machine learning capabilities to bring insights and intelligence into retailers’ businesses. Fashion was a category of particular attention as it is notable for its fragmented data sources and workflows..
Aristo AI brings together data sourced from across a business – including from the brand DNA documents through to design drawings and photo-shoots – and integrating this with a company’s Product Lifecycle Management (PLM) system before feeding it into a single unified system.
It is a similar story at Maeve AI, which also has a solution that plugs into a PLM as well as sourcing data from a fashion retailer’s sales and returns systems and then uses LLMs to identify errors in product designs before they move into the production process. These errors can include quality issues on certain materials, higher returns from specific suppliers and size issues with certain components.
Phoebe Gormley, founder of Maeve AI, says: “If you make a bad decision and then make products with it, then you have a problem. Once it’s baked into a product, then it’s the e-commerce teams who will have to deal with it and discount the products. These micro decisions can make a big difference.”
