Leveraging AI For Operational Efficiency: Walmart’s Strategy And Key Insights

One of the biggest retail giants, Walmart, has been at the forefront of leveraging AI in innovative ways to improve customer service, personalize shopping, and even increase associate productivity.

In fact for those of you who have read my book, “The Business Case for AI,” in chapter 2, I discuss how Walmart’s e-commerce search uses a suite of AI solutions aimed at achieving the following objectives:

(1) Maximizing the relevance of searched items

(2) Maximizing revenues while maintaining relevance

(3) Providing a seamless user experience

What I love about Walmart’s AI strategy is its purpose-driven nature. Instead of using AI for AI’s sake, there are many use cases where Walmart uses AI for true business benefits.

In this article, we’ll examine three other ways Walmart integrates AI into its operations and extract key learning points from each.

1. Chatbot to assist customers with their order

Purpose: reduce time to servicing customers, reduce employee workload

When customers have a problem with their order, they often want it sorted out immediately. To this end, Walmart has been using NLP/NLU chatbot technologies to answer customer questions related to the status of orders, process returns, and more.

Source: https://tech.walmart.com/

As a result of this initiative, Walmart has nixed millions of customer contacts, reverting the easier questions about order status and returns to its chatbot. This has allowed Walmart’s support associates to focus on the trickier service issues.

Furthermore, through customization of the chatbot technology, this chatbot handles localized languages in various countries, such as Mexico, Chile, and India, and also understands associated contexts and catalogs. Since customers are getting more immediate service, they saw customer satisfaction scores (CSAT) increase by 38%.

What can you learn from this customer service chatbot?

1. Simplicity & risk mitigation: As seen in the image below, instead of responding like your best buddy, Walmart’s chatbot provides fairly structured responses, limiting the chances of providing erroneous and superfluous responses. It sticks to the point, and the burden is primarily on understanding the customer’s question and assisting the customer instead of trying to sound human.
   
   Moreover, even if the chatbot misinterprets the customer’s question, its structured output reduces the likelihood of leading customers into undesired actions. Any mistakes in understanding the query would be more apparent to the user.

2. Build once, use repeatedly: Although this customer service chatbot was initially built for customers in the US, the company was able to customize it for localized uses throughout the world. Sharing the underlying chatbot technology means extreme cost savings for the company, as the work done for each localized situation is limited to fine-tuning and customizing the tech for localized uses. It’s much easier to maintain and improve a single chatbot technology instead of ten different ones.

Walmart’s customer service chatbot provides predictable responses

2. Generative AI-powered search function

Purpose: personalized results, and a more streamlined shopping experience

Walmart has implemented a generative AI-powered search function on its mobile app and website. This feature is designed to understand a customer’s search context and provide personalized search results.

For instance, if a parent is organizing a minions-themed birthday party for their child, they don’t have to perform multiple searches for minions-themed items. One for balloons, one for cake toppers, one for plates, etc., which is what they will have to do with regular keyword-based e-commerce searches. Instead, with this new GenAI experience, they can just pose the query, “Help me plan a minions-themed party for my daughter” and they’ll see products relevant to their goals.

Generative AI simplifies the retail search process, transforming shopping from scroll-driven to goal-driven. This makes the digital shopping experience more streamlined, potentially reducing the number of searches, clicks, and page navigation needed to find items. While maintaining revenues, this efficiency could even result in a revenue boost due to an improved shopping experience.

Walmart’s search results for “Help me plan a minions-themed birthday party for my daughter”

What you can learn from this GenAI-powered retail search:

1. Subtle but clever: Unlike many AI initiatives that try to change their user experience by introducing new chatbots in co-pilot style, which disrupts the customer’s old workflows, Walmart very subtly provides a chatbot-style experience in its search functionality. This means that customers can continue using their search function as they always have and reap these additional context-aware search benefits. ****

2. Incremental complexity: Walmart’s generative AI experience currently supports context-aware searches to provide personalized results. However, this capability can grow with time, allowing users to perform increasingly complex functions right from the search bar.
   
   Imagine being able to accurately place an order for several grocery items with a single search request that includes the quantities of those items. That’s a solvable use case. But, instead of releasing a firehose of such features, Walmart is incrementally increasing its GenAI capabilities, which is most likely to limit risks and ensure that trust remains high with customers.

3. AI-Powered Store Advisor

Purpose: waste reduction, and revenue loss prevention

Walmart is piloting a new AI solution aimed at minimizing food and fashion waste. This waste leads to millions of tons of unsold items each year, potentially ending up in landfills and costing Walmart billions in losses.

The initiative involves an in-store AI system that advises employees on managing product ripeness, seasonal fashion sales, and waste reduction strategies. For example, Walmart’s AI technology enables staff to scan items such as bananas to assess their ripeness. Subsequently, utilizing AI, a digital dashboard will propose actions for the product.

Depending on the product, the AI might suggest adjusting the price, returning the product to the vendor according to policy, or recommending donation as the optimal course of action based on the analysis. The AI tool helps employees make proactive decisions based on data insights, potentially reducing waste.

It’s poised for a pilot launch in Canada and has plans for global expansion. Walmart’s commitment to waste reduction aligns with its goal of eliminating operational waste across its North American operations by 2025, reflecting a broader industry trend towards sustainability and waste reduction.

What you can learn from this store advisor initiative:

1. Putting company-owned data to work: Walmart is able to build such a powerful tool not because it has special knowledge in AI techniques but because it has collected large volumes of relevant historical data. For example, to build such an AI advisor, at a bare minimum, Walmart would need historical data on product conditions, the actions that were taken for the products in those conditions, and the corresponding outcomes.
   
   Without such company-specific data, you can only get away with generic applications. And the interesting thing is that for a solution like this, you don’t even need sophisticated AI algorithms. Simple recommendation algorithms will work equally well so long as your data is reflective of the behaviors you want to mimic.

2. Boring but essential: While this AI advisor might not seem as flashy as a chatbot that covers a range of topics or one that attempts medical diagnosis, its role is vital in achieving sustainability objectives, significantly reducing waste, and boosting the company’s revenue. It’s a win-win-win on all fronts. These types of application opportunities often remain hidden until you start proactively finding employee and customer pain points and shortlisting those that are suitable for AI.

Final Thoughts…

Unlike many companies that use AI for AI’s sake, Walmart is a notable example of a company that applies AI where it truly matters while also adopting a risk-averse approach. The company progressively introduces more advanced AI features into existing AI-powered solutions and subtly incorporates AI technologies without causing disruption to the customer experience, as was seen in the chatbot and the generative AI-powered search example.

Importantly, as we saw in the store advisor example, Walmart creatively leverages its vast data resources to create powerful customized AI solutions. These solutions may address issues that don’t appear exciting on the surface but are actually crucial to the organization’s operations.

Now the question is, how are you going to take a strategic approach to AI integration? Which of the points outlined above will you attempt to mimic? Finding unknown but lucrative AI opportunities? Subtly integrating AI instead and progressively improving it? Significantly reducing the risks of generative AI mishaps, by taking creative countermeasures? Start with one.

That’s all for now! This article was originally posted here.