Making generative AI speak retail - Walmart and Kroger have big plans for LLMs

Grocery retailers have long made digital a core part of their operating model strategies, but now there’s a new tech topic on the shelves and it is, inevitably, AI. 

According to data from GroceryDoppio, a digital grocery intelligence resource created by Incisiv and Wynshop, between February and May, the percentage of grocers who discussed ChatGPT at senior level meetings rose from 67% to 83%.

But while generative AI in particular is a major talking point, how soon that translates into action is unclear. Only seven percent of grocery execs polled have actually invested in net new AI capabilities in 2023. The top challenges to adoption of AI are seen as budget (71%), proof of performance (69%), and infrastructure limits (63%), according to the data. That said, three-quarters (74%) of respondents are building AI into tech RFPs. 

More sophisticated personalization capabilities are what Kroger CEO Rodney McMullen expects to come from using AI tech: 

For years, Kroger has been at the forefront of using data and analytics, including Artificial Intelligence, to build a better customer and associate experience. By applying our data and AI-based personalization, we can better understand what truly matters to our customers and deliver more targeted and effective experiences.

As customers' digital engagement increases, we have new and more efficient channels to present the most relevant products and the right promotions at the right times, no matter where and how customers choose to shop with us.

As AI advances, we continue to work and constantly evaluate potential use cases throughout the business with privacy and responsible implementation in mind. For customers, we see opportunities to further simplify the digital experience and offer more accurate personalized recommendations. 

Kroger is putting its money and resources where its corporate mouth is, he adds: 

Our teams are working with search algorithms and generative AI to improve substitution accuracy and search results. Within our 84.51 customer research team [formerly dunnhumbyUSA] , we are already piloting several Large Language Models to summarize customer database sets. By applying AI to customer surveys and customer service logs, our team can analyze and categorize them in minutes versus days before. This allows the business to react to customer feedback more quickly and accurately, and then reflect these learnings in the customer experience.

McMullen concludes: 

Data is in our DNA. Our rich history as a technology leader gives us confidence that we will continue to effectively use AI, including more recent innovations. We also believe robust, accurate and diverse first-party data is critical to maximizing the impact of innovation and data science and AI. As a result, Kroger is well-positioned to successfully adopt these innovations and deliver a better customer and associate experience.

Meanwhile...

Over at Walmart, AI is also a core topic of interest, seeing as sitting alongside automation as being a transformative force over the next five years, as the firm disclosed in a recent investor meeting. 

The rise of generative AI is very significant, according to  Suresh Kumar, Walmart Global CTO and Chief Development Officer, but it’s only the most recent step in a long journey: 

We have been using ML [Machine Learning] and AI throughout the systems for a very long time. When you place an order online and if we want to do a smart substitution for you, the recommendation [of the replacement] actually comes from ML.

We use natural language in a lot of our internal applications as well as with customer-facing applications as well. A large fraction of the customer contacts are now automatically answered by our natural language processing systems, even internationally - Chile actually is one of the leaders out there. So, AI is deeply integrated into a lot of systems that we do.

But he admits that generative AI potential is very exciting for a company like Walmart, breaking down its impact to two main categories: 

One is around customer experience. We think that these capabilities are going to really accelerate the customer experience in a very meaningful way. It will help our customers be able to navigate our selection better, it will be able to help our customers go through their shopping mission, if you will. We have given some examples before of, ‘Hey, I’m looking for a cell phone for my daughter, what is the best cell phone that I want to find?’. For those kinds of shopping journeys, I think generative AI can have a really positive impact.

The other major area is internally around how our associates work, how we run our business, how we get insights from data. All of these are areas where I think, as generative AI becomes more and more mature, we will start seeing the impact of that. And we have already started a number of initiatives where we are pulling in the latest developments that are happening outside, both for our internal tools and to drive customer experience.

Walmart’s strategy will be to build its own LLMs as well as to partner with technology firms to leverage generative AI, he adds: 

There are a lot of foundational breakthroughs that have already happened. Whether it’s OpenAI, whether it is on Google's side, or even open-source, we want to take advantage of all of the breakthroughs that are happening. But we want to do that in such a way that we take that and we make it uniquely Walmart, so that it works on Walmart data, speaks retail. For those kinds of things, we will take the developments outside, and we will make it work within the Walmart environment in a way that is going to be beneficial for our associates and for our customers in a unique way.

Going back to that Grocery Doppio data, grocery firm execs who were polled predict AI will be used to augment (43%), automate (39%) or replace (18%) grocery staff. From Walmart’s perspective, Chief People Officer Donna Morris, AI can indeed be a transformative influence internally: 

A lot of our investments in technology have been focused on our customer, and frankly, the ability to serve our customers. And for the most part, a lot of our technology was aimed at our very large frontline associate base, which it should be. 

But when we think of the early opportunities with generative AI, it actually will change and shape the ways of working for those that are our campus office associates, and optimize their performance, just like we’ve been leaning in to optimize and make our frontline more effective and efficient.

Walmart is taking a people-led approach to its thinking, she adds: 

We can actually make our people better by augmenting their work through generative AI. I think there’s incredible opportunities across each of the functions to actually leverage [this]. From a people space, in general, I think there’s a disruption, because we’ve long talked about the need for self-service. Most individuals just want to find things on their own. They don’t want to call someone. They want to find it. And so, in a world of generative AI, it’s opening up the opportunity for individuals to self-serve. 

I think you’re going to see some pretty exciting things that we can do to support our campus office associates being as effective just like we’re making the jobs in the front line better because of technology.

My take

Back to the GroceryDoppio data again - it is interesting to note that while 82% of respondents say that adoping AI will be a necessity to compete in the future, only 13% of them plan to spend non-budgeted funds to explore AI solutions this year, while only 11% reckon AI will significantly impact staffing over the next two years. The revolution may a little while coming…