Bill Groves, Former Chief Data Officer, Walmart, speaks with Riddhiman Das, CEO, TripleBlind, about the ideal approach to data safety and Walmart’s data strategy in general.

Speaking on the driving motivation for implementing AI in Walmart’s strategy, Groves says that like most other industries, Walmart considers AI a primary disruptor. The organization sees it as a capability that can be used to deliver better services to customers. “That’s the way you can be more efficient as well as drive more revenue. There’s a lot of different things you can do with AI and we see it as being a primary game changer for us,” he says.

However, AI systems need data for training, fine tuning, improvement, and building. Groves sheds light on the challenges there, and the approach to leveraging first-party and third-party data for training and leveraging AI strategy.

He emphasizes that to make decisions based on AI looking at insights with data, the data has to be high quality. “It has to be a golden asset and has to be managed and governed. That’s the key. There are challenges for every company, not just Walmart, when it comes to that data. There are challenges with first-party data. Typically, you run into challenges around duplicate data, data quality, conflicting data assets sometimes.”

To sort this out, Walmart conducted an exercise over three years to build up the data asset and make it a golden asset. It’s a larger challenge in the AI and ML space because of the volume and types of data matter.

“In the first-party space, one of the things we do have is a lot of data. We have a lot of data we can use to make decisions and build AI and ML solutions. That being said, third-party data enrich the decisions that you’re making,” Groves explains.

He points out that some third-party data doesn’t have a heightened sense of review to ensure usage and validity from a regulatory perspective. “A lot of third-party data is very easy to use and is very commonly used in the industry by many companies. We focus heavily on the first-party data but we also do look at third-party data,” Groves says.

Regarding the challenges of using first-party global data from a regulatory perspective, he points toward a couple of aspects.

First, Walmart strives to be the most trusted retailer, and it is considered a competitive advantage as it approaches everything from a data-AI perspective. This starts with a few questions. “One is, is the customer at the center of this? Are we making a decision for delighting our customers or for the customer’s benefit? And we hold their trust sacrosanct. We will never defy that,” Groves says.

Second is the regulatory aspect. The guiding principles have to be kept in mind while looking at different data sets. For example, customers in the United States and the customers in Asia act very differently. “So we typically keep them more regionalized,” Groves says. “We don’t pull the data across because one, there are legal issues with that, and also, it doesn’t make a lot of sense from a customer perspective.”

Walmart will never look at customers individually but will instead look at them at a more aggregate level, he adds.

“Say there were a thousand turkeys sold in Texas and 5,000 in Oklahoma. So, we need to put more turkeys in Oklahoma. That’s how we look at data, and that’s how we work with it,” he concludes.