The Future of Artificial Intelligence in Retail

Reposted from TotalRetail

By Andy Narayanan, Vice President of Intelligent Commerce at Sentient Technologies

The uses of artificial intelligence (AI) that get the most press are usually the big, splashy ones. Whether it’s IBM’s Watson beating Ken Jennings at Jeopardy, DeepMind besting Lee Sedol at Go, the massive influx of news about self-driving cars, the growing personal marketplace, or Elon Musk’s increasingly public trepidation, these kinds of AI stories have a way of capturing public attention. But quietly, AI powers search and recommendation engines at places like Google and Netflix, filters out obscene images on your favorite social networks, and proves complex mathematical theorems. It is, increasingly, everywhere.

You probably hear far less about AI applications in retail. On their face, they aren’t as sexy as beating a champion at the world’s oldest (and most complex) board game or writing rudimentary piano songs. However, AI in retail is something that will affect everyone who shops online in the coming years. Which is to say, well, pretty much everyone.

So when will you see AI in retail? Unlike self-driving cars or any of the other loud announcements we mentioned above, AI in e-commerce is coming about more subtly. But make no mistake: it’s coming. After all, the ingredients are there. Retailers have the data that drives successful AI and the bottom-line reasons for embracing its transformative possibility. In fact, many retailers are already doing this. Image recognition systems, product recommendations and chat functionality are already being powered, at least in part, by AI systems.

The pertinent question isn’t necessarily when but where you’ll see AI deployed. And the truth of the matter is that AI can benefit essentially every step and process of e-commerce, from site layout to personalization to — and this part is extremely important — customer happiness.

One of the big promises of AI is a truly personal experience. Right now, if you head over to your favorite site, chances are it wouldn’t look a lot different from how it would look if I visited. But we’re different people, with different preferences, different likes and dislikes, different in-the-moment needs. Sure, you might see some level of personalization — e.g., a gentle reminder that you looked at a product a few weeks back or recommendations based on users like you — but that’s not truly personal. It’s based on your past purchases and walling you into a certain user group.

AI, on the other hand, can be truly personal. If you search today for, say, “red shoes,” you and I will get the same page of top results, page two results, and on down the line. But with AI, you won’t see this. Imagine instead clicking on a pair of red shoes and being presented similar options, not based on language, but based on the image itself. Instead of seeing a page of self-same results, you see a curated catalog, all based on the exact pair of shoes you clicked on. AI can look at the subtle shade of red you chose, if there’s a fringe or flourishes, toe type, or any other subtle similarities. The more interesting part, though, happens next. With each subsequent click, the AI learns what it is you’re interested in. It creates threads through your choices. It understands what you want, right then, in the moment.

This is real personalization. It lets individual users explore catalogs how they want to. It lets you shop your way and me shop my way. And these are the precise experiences we’re missing in retail now.

With AI, retail can — and should — change more than that. It shouldn’t just help us after we search for something, but before it. The first action a user takes on a site should transform the entire site to that user’s preferences. It’s all well and good if a site sells everything from washing machines to socks, but if you’re just there for socks, you probably don’t care about anything else. That doesn’t preclude you from, say, shopping for washing machines afterward, but once the site learns what you’re looking for, it should reconstruct itself around the experience you need, right then.

The site should maintain its branding and core elements, but the attributes should change right then. Right now, shopping online is a lot like heading into a Costco. Sure, all the products are there, but it’s up to you find them. There’s nobody helping you, walking you down the aisle, then heading to the storeroom to find exactly what you want. It’s just product. Lots of product, sure, but that can pose its own problem — the famous “paradox of choice.” It’s overwhelming. AI should allow you to, in essence, zoom into a certain part of a certain aisle and effectively ignore the rest of the aisles with their endless distractions. From there, you should begin with the immersive experience we outlined above: simply browsing through images of products. Every action you take trains the AI to help you find exactly the thing in exactly the category that fits exactly your unique needs and sense of style. After all, having thousands or millions of products isn’t worth much if a shopper can’t find the one they actually want.

E-commerce companies that embrace AI at every level will provide better experiences at every level of the shopping journey. In places, those differences will feel massive. In others, they will be subtly differentiated. But AI will change the storefront completely. Gone will be generic sites that look the same for you as they do for anyone else. With AI, that will be replaced with stores, whole malls, where every choice and action is tailored to your action in the moment. You won’t have to pull product choices from endless catalogs. They’ll come to you, based on just the simple, intuitive actions you already make.

It won’t be like wandering a big-box store anymore. Every store and shopping experience will be hyperpersonal, from the moment you land on the site to the moment you check out. And it’s already happening. The retailers that embrace this now will be a step ahead. Make no mistake: AI is already powering your movie recommendations, your favorite social networks and your search engine. This is just the beginning.