Walmart AI Patent Signals Continued Pressure to Compete with Amazon

Walmart brings artificial intelligence to brick-and-mortar stores.

The retail giant has filed a patent application for a “machine learning framework.” Walmart’s patent describes an infrastructure and method for transferring data between machine learning models, which is intended to make data collection easier and save time and resources.

“While the use of machine learning algorithms can be beneficial to organizations, the upfront costs can be very high for a number of reasons,” Walmart said in the filing.

Walmart’s filing details how to transfer “feature data” between the two models. The term feature data has a variety of potential meanings in Walmart contexts, although when applied to customers, the filing notes that it can include things like interactions, demographic data, or “identity mapping data.”

The technology itself is relatively simple: Walmart’s system uses a “configuration file” to determine exactly which features, along with a subset of related historical transaction data, are relevant and necessary to train a machine-learning algorithm.

That data is then placed into an “output file” that is “configured to be passed between two or more machine learning models.” Because these output files can be easily created and used to train several different models, the framework allows Walmart to more easily develop and deploy models for purposes such as customer goal setting or predictive analytics.

This isn’t the first time Walmart has explored AI. The company has previously applied for patents for ways to predict user acquisition and churn, as well as for a personalized recommendation engine based on a user’s transaction history. It has also tested AI in a store to advise employees on how to sell products, and launched an AI-powered logistics offering earlier this year.

While this may seem like a sign that the retail industry is trying to keep up with the AI-powered technology craze, smaller brick-and-mortar stores appear to be in no hurry to innovate, said Thomas Randall, consulting director at Info-Tech Research Group.

“I think for smaller shops, there’s less fear of missing out. They’re just trying to maintain their profit levels in the face of the unstable supply chain that it is,” he said. “Larger institutions … are still trying to figure out the right use cases for themselves. I think it’s fair to be a little bit behind.”

A likely factor in Walmart’s push, however, is Amazon. Walmart has long battled competition from the e-commerce giant, beefing up its third-party online marketplace, launching its own subscription offering to compete with Prime and even testing its own drone delivery service.

That competition extends to technology strategy: Walmart has in recent years pulled back from Big Tech’s cloud offerings, building its own internal server infrastructure to handle processing. Building its own machine learning development infrastructure, as the patent details, is likely part of that push for independence, Randall said.

“Because of the competitive environment, (Walmart) has to bring in a lot of its own technology to … continue to compete with other large e-commerce retailers like Amazon, who obviously have their whole ecosystem of AI infrastructure, core models and partnerships,” Randall said.