Inside ALDO’s in-house generative AI and machine learning strategy

“Having machine learning and generative AI is something that people really are working on,” he said. “But all of them require to have a good foundation, from the data perspective, to be able to consolidate all of [the functionalities].”

Currently, ALDO is using everything from machine learning to generative AI, including predictive AI, which uses machine learning to make predictions about future events. Like other companies, generative AI is in nascent stages, used to generate text, SEO and product descriptions while predictive AI — a work in progress, he says — is used for things including demand and sales forecasting and discount optimization. 

It takes data to build and train these machine learning and generative AI capabilities, but ALDO’s data efforts go back to at least the last five years, an aggregation of customer patterns like website clicks and in-store purchases, Nayebi said. To ensure that data’s safety, the retailer has its own data clean room that aggregates insights that “could apply to everyone” as opposed to personal identifying information, he added.

“We built all of those foundations to do all of this data and AI products,” he said. “We are already having our retail e-comm supply chain running based on that data. Now we’re leveraging the same data for recommendations for demand forecasting and other insights.”

To put it plainly, the data from ALDO’s store interactions and e-commerce business is brought in, anonymized and consolidated to be used in AI and machine learning models including for product recommendations, demand forecasting and sale forecasting, he said. The company declined to share specific figures to show AI’s impact on the business but expects savings in marketing spend as things progress, per Nayebi.

As data privacy initiatives continue, no matter how slow the death of Google’s third-party cookie is, ALDO is aiming to “future proof” its data sets, relying on its own first-party data to feed the so-called AI beast.

“We know that the tracking and all of those types of things is not going to be something that we should rely on in the future,” he said. “So what we try to achieve is really to bring insights based on the aggregated level of customer patterns.”

In the generative AI boom, some companies have been looking to AI for quick headline grabs while others are looking to it to solve efficiencies — something AI has the potential to solve. However, it remains to be seen whether AI is the workhorse that will deliver.

Every company is feeling some pressure to figure out how to incorporate generative AI and machine learning into their processes, said Brian Yamada, chief innovation officer at VML. The question is how much does a company commit to building internal systems for the future of AI versus working with external partners, like OpenAI or Microsoft AI, he added. 

“That’s going to be a mix of build-buy. The rate of speed that the market is moving is making that problem. It’s a big challenge because it’s hard to keep up,” he said. “I feel like I’m spending my time drinking from the AI fire hose and trying not to drown.”

Freddy Dabaghi, managing director of activation at Crispin ad agency, made a similar comment.

“Brands are leveraging their own tools and systems to have more control over data privacy. These internal teams are also building out AI task forces, but they are not running at the speed of marketers,” Dabaghi said in an email.

The key, Yamada added, is being flexible as AI shakes itself out, revealing its most efficient use cases as the hype cycle continues. Building out internal systems to rake in data, train models and create internal processes, like ALDO is doing, could be a step in the right direction even as the landscape stretches and changes, he added. 

“The key is just being flexible, not getting too locked in or too rigid in a way,” he said, regarding the future of generative AI. “We’re moving into much more fluid environments.”