Blog Home

Retail AI: It’s Not Just for the Fortune 500 Anymore

A new study from IDC predicts retailers will spend $5.9 billion on AI tools like automated customer service and product recommendations this year alone. And 79% of retail and consumer products companies expect to use intelligent automation to learn more about their customers by 2021.

Yes, retailers are making big investments in machine learning and automation, and the returns so far are strong—but they are unevenly distributed. The biggest beneficiaries of these trends are the already-powerful retail giants like Amazon, Walmart or Target. That’s because these powerhouses have 1) the largest pools of consumer data to feed their models, and 2) the resources to build, test and deploy AI at scale.

That head start has given the early leaders an almost unfair advantage, and they’re not likely to relinquish it anytime soon. However, thanks to some innovative new services that are now emerging, smaller companies (or larger ones who are late to the AI revolution) will be able to harness that power as well.

What is AI?

Here’s a decent working definition: ‘Artificial intelligence’ describes machine systems that mimic cognitive functions we normally think of as human, such as learning and problem solving. On a practical level, it can be anything that helps businesses work through large piles of data that used to be analyzed manually.

The value of any AI to retailers boils down to three questions: Does it provide new information or opportunities that otherwise would take manual work to identify, analyze and implement? Does it do so faster than I could do manually? And ultimately, does it provide more value to my customers?


Some examples

AI is already deeply embedded in many key retail functions, such as:

Personalization

Recommending products based on past purchases (or some other data in a customer profile) is widespread. Now, thanks to advances in facial recognition, AI-driven personalization is starting to be seamlessly incorporated into brick and mortar retail as well. Retailers are experimenting with the ability to either recognize individual customers in stores as part of a loyalty program, or even respond to emotional cues on customers’ faces to provide better service.

Price optimization

Setting prices used to be a slow, inexact, manual process. Getting useful customer feedback could take months. Now companies like Amazon adjust their prices as often as every 10 minutes. And there are SaaS tools that can determine optimal prices based on the rest of the market, or look at your customers’ past purchases to help find that sweet spot between the price they’re willing to pay and the price that nets you a healthy profit margin.

These are powerful uses of AI, but what is most exciting is the expanding availability of retail AI technology. Effective AI tools are now available even to small or mid-sized retailers. Along with the pricing engines  mentioned above, there are tech firms offering a wide range of AI-driven business services, from automated analysis of focus group research, to AI-created product videos and many others.


Let’s use our powers for good

Any tool has the potential for abuse or misuse, and AI is certainly no exception. Consider how AI can make dark UX patterns in ecommerce even more efficient, as when travel booking sites stimulate a customer’s fear of missing out by displaying a running count of how many times this specific hotel has been booked. The aim of these techniques is to make the user click a button, without caring too much if it’s really the right hotel for them. This may increase short-term profits, but does nothing to create long-term value by enhancing the customer experience.

Another danger of AI is the core truth that it can only learn from the data set it’s been given, and only in the ways it’s been programmed to learn through its algorithms. Thus, AI will tend to replicate the biases of the team that created it. An obvious example is the failure of facial recognition and autonomous driving systems to recognize and respond to nonwhite people. Only truly diverse teams, guided by strong ethics, can build and implement AI systems that benefit everyone. With AI, diversity and ethics need to be built in from the ground up, not added in as an afterthought.

AI is undeniably part of the future of retail commerce. As long as we watch out for—and work to avoid—its dark sides, and work to ensure that its benefits are spread widely, we can help to make it a future we look forward to with anticipation rather than dread.


Editors' note: A version of this story was first published by TotalRetail.


comments powered by Disqus