The Future Is Now: How AI Impacts Today’s World of Retail

Published in

Latitude

8 min readDec 22, 2019

If we looked at it from the standpoint of a human lifespan, AI (artificial intelligence) is currently crawling its way through infancy. But, as it turns out, AI is no ordinary child. Although it has a lot of growing up to do, AI is already changing our lives in everyday ways we might not even realize, especially in the world of retail. For this article, we’ll focus specifically on how retailers are using AI to elevate the user experience (UX).

For an infant, AI is a pretty active kid. And plenty of companies are putting that energy to use.

But First, What Is AI?

Using AI effectively begins with understanding what it is at its core. AI is any computer program with the computing power and ability to perform tasks that would otherwise require human intelligence. With the increase of computing power and available data, AI use exploded. What caused the increase in data? We have our friends the internet and social media to thank for that.

So which functions do AI serve?

As of now, most AI used in retail (or in general) fall into the category of “weak AI.” Siri and Alexa (America’s sweethearts) fit the bill here. In retail, these AI currently fulfill three basic functions: personalization, prediction, and communication.

AI, for instance, could help manage inventory and notify in-store staff when certain items need refilling, as they do at Walmart or Lowe’s. AI could also answer questions from online shoppers through chatbots. One AI we’re all familiar with is the AI employed by Facebook and Instagram that uses an algorithm and machine learning to figure out what ads or content you’d be most interested in seeing based on your personal preferences and search history. Even though we consider these AI “weak,” they have incredible strengths and potential.

Here’s how retailers are tapping into the strengths of these “weak AI.”

AI & Coffee — Starbucks

What do AI and coffee have in common? The answer — Starbucks.

As they did earlier, Starbucks once again changed the way customers experience coffee. This time, they did it through an AI embedded in their mobile app.

With over 17 million users, the Starbucks mobile app collects a fair amount of data. Every time a customer uses the app to collect rewards, search for a store, or place an order, the Starbucks AI uses this information to personalize the customer’s experience later on. Data collected in this way allows Starbucks to notify its baristas of the customer’s preferred order, even if they’re visiting a store for the first time. The Starbucks AI can suggest new drinks to a customer based on their previous preferences, the weather, holidays, or time of year. To top off the experience, the Starbucks AI even allows customers to order drinks through the app using written text or voice messaging. Not bad for a “weak AI.”

Beyond its mobile app, Starbucks uses AI to solve a variety of other challenges. By analyzing data on location, traffic, or the demographics of a neighborhood, Atlas, a mapping AI developed by Esri, helps Starbucks decision-makers decide where to put new stores. Atlas even accounts for the location of other Starbucks stores to help new shops avoid cannibalizing the business of existing ones. At Starbucks, both strong coffee and weak AI get the job done.

AI & Skincare — Olay

Figuring out the best skincare regime can be hard work, but AI can drastically simplify the process. Since 2016, the Proctor & Gamble-owned skincare company, Olay, used its AI, Olay Skin Advisor, to assist customers in picking products that worked best for them. By looking at a simple customer selfie, Olay Skin Advisor gains all the information it needs to recommend a fully comprehensive skincare routine. Of course, the AI also recommends Olay products that can help the customer meet their personal skincare needs and concerns. So, what impact did AI have on Olay’s business so far?

According to Olay, the company doubled its conversion rates since launching Olay Skin Advisor. Beyond that, the AI boosted the online shopping cart size of each customer, especially in China, where customers increased their basket size by an average of 40%. Having reached over 4 million shoppers, the AI allowed Olay to reduce its online bounce-rate to a third of what it was prior to AI use. Olay also reports that 94% of shoppers who used the AI found that the AI’s recommendations were right for them and met their needs. With AI, even a selfie can offer information a company needs to boost its sales and better serve its customers’ needs.

AI & Outdoor Goods — The North Face

Ever experience analysis paralysis — the uncomfortable feeling of having too many choices and not knowing which option to pick? Shopping for a garment that you need to perform especially well under harsh conditions — like winter weather — can make the anxiety over choosing the best option even worse. Down or synthetic insulation? Midweight or heavyweight? Hooded or hoodless? If only someone could just pick the perfect jacket for you. If you shop at The North Face, that someone might just be an AI named Watson.

Watson, developed by IBM, provides customers at The North Face with a cure to analysis paralysis. In an effort to bring the knowledge and support of an in-store retail expert to the online shopping experience, The North Face and IBM programmed Watson tell a customer which jacket to buy based on how the customer plans on using the jacket. To receive Watson’s recommendation, a customer would answer a number of prompts from Watson on The North Face’s website, such as, “What’s the weather like in the area you plan on using the jacket?” From there, Watson compiles the data and suggests several recommendations based on the customer’s information.

How Does Watson Do in Reality?

It’s a novel idea that theoretically creates an easier online shopping experience. But does Watson work well in reality? That’s another thing entirely.

According to one author from the Harvard Business School, Watson didn’t always make useful recommendations. At times, the AI suggested a coat that wasn’t in stock or that didn’t come in the customer’s size. Senior Director of Ecommerce at The North Face, Cal Bouchard, compared Watson’s early level of artificial intelligence to that of a second or third-grader, which means customers shouldn’t expect complete accuracy every time. As Watson collected more information, it would learn from customers and theoretically improve at offering recommendations.

Even if Watson provided accurate recommendations in most cases, most customers still prefer to try on especially pricey items, like a jacket, in-person. That’s something AI just can’t compete with, even if they can make an accurate recommendation. And, as of late 2019, The North Face website doesn’t include an option to search for jackets using Watson. For now, it looks like the AI has some learning to do on its own time.

The story of Watson reveals a lot about AI. The truth is, AI technology has not only limits but also times where it completely drops the ball and fails pretty hard — even for big retailers like Amazon.

AI Flops & Fails

For as much success as Amazon has enjoyed as a retail innovator, the company has experienced a few flops while experimenting with AI. In one case, Amazon developed an AI to screen and choose the best applicants for software engineering positions. However, the AI’s programmers soon noticed something about the AI — it had a bias towards choosing white men.

Why? Well, as part of its machine learning process, the AI analyzed current Amazon software engineers to establish a benchmark. And, since many of the benchmarked employees were white men, the AI determined that white men must make the best software engineers.

That’s not the end of Amazon’s biased AIs either. In 2018, the American Civil Liberties Union (ACLU) reported that out of all the faces Amazon’s AI, Rekognition, failed to recognize, 39% were people of color, even though people of color only make up 20% of Congress. In another study by MIT and the University of Toronto, researchers found that each facial recognition system that they tested had a 1-in-3 failure rate while trying to identify women with dark skin.

So, What Do AI Failures Say About AI?

Perhaps AI gaffs provide more of a commentary on the failure of the people, companies, and systems that invent the AI rather than the AI themselves. Even as machines learn on their own, it’s ultimately people that program them. And people are prone to prejudice, mistakes, and shortsightedness. To make AI an innovation that truly benefits the human experience, we need to keep in mind our own human tendencies towards making errors and bias decisions.

The Takeaway

AI has undoubtedly changed the world already and it will continue to do so. How it changes the world depends not on machine learning and AI performance but on the humans and the intention behind the AI themselves. Using AI to successfully enhance UX will require constant testing. It will also require people to question the systems in which AI operate, disrupt, and support.

As with any other tool, the impact of AI depends entirely on how we use it. Will we use AI to enhance the human experience and help us process data to make more informed and wise decisions? Or will we use AI to cement current systems of the past that keep us trapped and stagnant?

The answer depends entirely on us, the questions we ask, and how mindfully we approach the opportunity of AI.

Want to learn more? Feel free to reach out at any time. We would love to chat!

The above piece was written by the Latitude Supercharge Research Team, which includes Connor Beck and Carter Jensen

Supporting Resources —

Resources and original reporting of the above points covered by the following publications — Emerj.com, Insider Trends, Tech Republic, lexalytics.com, Harvard Business School, Venture Beat.com, BuzzRobot, Hackermoon, Forbes

At Latitude, we love taking incredible brands of all sizes and elevating them through tech-fueled experiences that add true value. From pop-up retail to permanent build-outs, our team brings brand stories and modern-day commerce together to truly stand out. Want to learn more? See our case studies. Give us a shout.