Aldo’s AI Investments Include Celebrity Picks and Demand Forecasting
Aldo Group is helping develop a generative AI model that lets shoppers receive personalized product recommendations based on a particular celebrity.
For example, a shopper can request, “I need to look like David Beckham,” and the model will generate an image and recommend shoes in a conversational way, according to Matthieu Houle, CIO of Aldo Group.
The footwear brand and retailer partnered with McGill University in Montreal last fall for a hackathon to build retail solutions using genAI. The model was developed by a team of students and is currently being brought into production by Aldo.
Delphine Demand Forecasting
As part of its artificial intelligence exploration, Aldo has also developed a demand forecasting machine learning model known as “Delphine” to help ensure inventory is available regardless of the channel, from in-store to TikTok Shop.
“We gave her a name because we just want to make sure that people don't think that she's going to have all of the future [responsibilities],” Houle explained during a virtual event hosted by SAP last week. “It’s art and science, so we have the fashionistas and buyers that understand the human part of planning. We just want Delphine to be at the table when they do their plans because she's very good at crunching data.”
Partially subsidized by the government in Canada, Delphine also intends to provide employees with more time to spend on such creative tasks as fashion trend predictions. While the model can’t predict trends, it can provide employees with an analytical starting point, said the CIO.
Aldo has a broad partnership with SAP and leverages the provider’s composable systems for both front- and back-office functions, including ERP, supply chain, and e-commerce. Delphine uses data from both Aldo data sources and SAP applications, and success has required serious time investments in data management, governance, and integrity, Houle stressed.
“Even if it's the same data, sometimes the data, and the way you need to manage it, is not the same as when you do your Sunday sales report. So we've invested quite a bit of time on this,” he said. “Pulling [data] out of SAP and making sure it goes back into our workflows and processes is an area that's not as sexy as the AI itself, but an AI by itself doesn't do anything if people are not able to use it.”
Still, managing the data flow between the front and back offices has been fairly straightforward, he said, as the back offices primarily handle product information, inventory, and pricing, while the front manages consumer data.
One of their biggest hurdles has rather been shadow processes and hidden data, such as what’s often found in spreadsheets.
“That's our biggest problem, truly, because then you have supply chain having a vision of the truth that doesn't match what finance sees,” he said.
Employee excitement surrounding the Delphine model has helped bring such hidden data to light, per Houle, and the company successfully integrated the recently acquired Sperry brand into its tech stack in just three months, including product wholesale, e-commerce, and store systems.
Data Ownership
Like many consumer goods companies exploring generative AI investments, Aldo prioritizes the business value that a potential use case can deliver. With the case of Delphine, it was determined that the model could add at least one point of gross margin, Houle said.
While technology investment decisions are made in collaboration across the business, data and artificial intelligence reside under Houle, which aligns with their software ownership structure.
“If you have data on one side, and the way to use it on the other side, you just duplicate the needs for leaders,” he noted, and can result in having too many cooks in the kitchen.
Aldo’s head of data and AI reports into Houle, though the company also has data owners within the business side of supply chain.
“I think it's better if it’s under one roof, in my opinion,” he said. “But again, we're [small]. It probably changes depending on your size."
Looking to stay connected to the consumer goods industry? Sign up for a new kind of insider access with CGT's texting community. Share your opinion, and send your questions right to Lisa Johnston for 1:1 conversations. Texts arrive twice a week. Learn more or sign up below.
