'Voice assistants, augmented reality will transform retail'

"Voice assistants like SIRI and Google Assistant, and AR/VR (augmented reality, virtual reality) will drive the next large shifts in the retail industry," said Suresh Kumar, Executive VP, Global CTO and Chief Development Officer of Walmart.

Augmented reality will become a natural way for customers to experience items they want to try and purchase, Kumar said. Voice assistants—that combine voice recognition, language processing algorithms, and voice synthesis—will have applications for customers, as well as for retail employees in tasks like locating products faster, he added.

The 56-year-old CTO of Walmart highlighted the strides made by the $559.2 billion retail giant in the past couple of years.

“We started by making sure we have the infrastructure that allows for rapid innovation and rapid testing—at scale," Kumar told Shereen Bhan, Managing Editor of CNBC-TV18, at Converge@Walmart, a virtual conference on retail technology, organised last week by Walmart Global Tech India.

"Even before COVID hit, we had started a journey of building a technology stack that is going to allow us to innovate,” he said. It began with a hybrid cloud infrastructure to harnesses the best of public cloud and Walmart's private data centres, he added.

"The hybrid cloud allows us to harness all this data and reason over it by using ML, (and) advanced analytics," Kumar said. "We are trying to invest in those areas where we have an outsized impact, while we continue to modernise and drive innovation."

Kumar joined Walmart in 2019, and has previously worked with tech giants Google, Microsoft and Amazon. He highlighted the role of data in faster decision-making at Walmart.

“The data we get, from running our supply chain and transportation systems, to even understanding how our customers interact with us and what they care about, helps fuel the customer experience,” Kumar asserted.

He illustrated technologies-enabled efficiencies by citing a machine learning (ML) model that helped Walmart optimise pricing and timing for Walmart's apparel division, resulting in $30 million of savings and faster replenishment of in-season apparel.

“We built an ML model to understand the best way to be able to drive markdowns in as localised a way as possible,” he said.

The ML model factored in variables like locations across the United States, climate by region, and customer preferences. "As a retailer, we need to make sure that we optimise our price. When we need to make space, we need to move our existing inventory. Typically, this process is laborious and manual," Kumar explained.

“Human beings are not able to optimise more than two or three variables, but ML systems are optimising millions of variables,” he asserted.

Kumar's appointment at Walmart has led to better integration of the company's brick-and-mortar presence with its online assets across geographies.

Last year, the company introduced Walmart+ in the US market, a membership offering with omni-channel shopping benefits, such as unlimited free shipping on select items, unlimited delivery from stores, fuel discounts, and mobile scan-and-go for in-store shoppers.

Kumar has also pushed for Walmart Global Tech working faster and collaboratively across different regions, including India, to cross-pollinate learnings of one geography with other relevant markets or businesses. Walmart Global Tech India has more than 7,500 employees.

"The underlying technology (in any market) has got a lot of common elements to it," Kumar said. "We have tried to build the common building blocks, and then customise solutions to the specific business problem that we are trying to solve in a market," he said.

"At the core, we are really about being able to help people save money and lead better lives," Kumar said. "To do that, we have to understand what their needs and what they care about. We make sure that our business—whether it is supply chain or merchandising or associate experience—is optimised to be able to serve that mission properly."

ALSO READ

We have to think local, and execute locally: McMillon