How AI Will Make Commerce As Natural As Talking To A Friend

Im cooking dinner, and Alexa tells me that two lightbulbs in the kitchen need to be replaced soon. She asks, “Would you like to reorder your last purchase of high-efficiency light bulbs, and have them delivered in the next few days?” I say yes, and avoid cooking in the dark. This interaction, which is possible today, is only a glimpse into what the intersection of voice recognition technology, artificial intelligence (AI) and predictive retail can make possible.

Retailers are learning to better anticipate customers’ needs and ultimately provide smart, personalized experiences based on data. Fueling this evolution in commerce are three core AI-related technologies:

1. Voice Recognition

Mary Meeker’s latest trend report predicts that voice will lead to a reimagining of how we communicate and interact with machines. We are already seeing the wheels in motion: 58 percent of mobile searches are made via voice, and we’re just in the early stages of adoption! Talking to your device to turn lights on, play music or set timers is now standard with the rise of Alexa, Cortana and Siri — and it’s becoming mainstream for shopping too. Like Alec Baldwin, you can reorder your favorite products like Bresciani cashmere socks with a few simple words. Or thanks to a Starbucks integration, you can pre-order a latte through Alexa on your way out the door.

We haven’t yet reached the golden age of voice recognition tech. Today there are small but persistent issues with devices having trouble distinguishing between a user and other voices in a song or commercial. Since its launch, users have noted frequent instances of Alexa accidentally taking commands from a TV commercial, mistaking it for the user’s voice.

There are larger issues at hand too. We need natural language processing (NLP) to become sophisticated enough to understand us regardless of whether we mumble, talk quickly, have a strong accent, or essentially speak like a human instead of a robot. We need to find a way to use this technology to support less transactional activity so it can help with discovery and inspiration. And from a brand perspective, it will be important to ensure that the voice interface doesn’t homogenize all experiences so the consumer can’t differentiate them or gets bored.

2. Predictive Retail Driven By Smart Data

Today, GE’s smart dishwashers, laundry machines and other home appliances automatically order your household supplies when they’re running low. Imagine this level of automation applied to every purchase, like auto-refill sunscreen based on usage patterns. Your hiking shoes remind you when you’ve logged 600 miles and should buy a new pair, and your raincoat detects when it’s losing its water resistance and needs to be replaced.

Soon, machine learning will lead a shift from this type of responsive replenishment to predictive commerce. Predictive retail is the ultimate in personalization, based on subtle patterns detected from the intersection of all the data available. Besides purchase history, product preferences, and usage patterns, retailers can take cues from other data sources such as social profiles, browsing behavior, geolocation, weather, calendar events, pricing and inventory fluctuations to proactively surface recommendations. Imagine that you have a wedding to attend in Hawaii in a month, and you get a curated list of 10 items to bring with you in your size, appropriate to the location and occasion, tuned to your tastes.

The two main barriers to making predictive retail a reality are sophisticated AI to crunch the masses of data into real-time, actionable insights on an individual shopper level, and pure data hygiene. Investors are supporting this evolving market without hesitation, putting an unprecedented $1.05B into artificial intelligence startups in a single quarter last year. We are at a tipping point with the influx in funding, developer interest and retail demand, so expect this to come to fruition sooner than you might expect.

3. Conversational Commerce

Bots are powerful communication channels which meet customers on the channels they’re already using, reduce the need to call customer support, and use contextual information to inform a more productive conversation. The North Face created a digital shopping tool powered by Watson that asks shoppers a series of questions related to the coat they’re shopping for, like where and when they plan to use it. The bot then makes a recommendation based on what users with similar responses have purchased and reviewed positively. Shoppers who use the tool make a purchase more often than those who don’t.

This progress aside, chatbots are no doubt in the “trough of disillusionment” per Gartner’s infamous hype cycle. 2016 was a year of inflated expectations about the potential of chatbots, especially after Facebook announced its beta Messenger platform for chatbots earlier in the year. Everlane, which was one of Messenger’s launch partners, recently discontinued their bot on the platform, although other retailers are experiencing success with the channel. But just because bots haven’t lived up to the hype doesn’t mean they are a flash in the pan. By having a single, clear purpose before evolving to tackle more sophisticated activities, bots can provide high consumer value with low friction, which will lead to increased adoption.

Both established and emerging retailers are already testing the waters with the first wave of AI-powered technology. And with the AI industry expected to be valued at $16 billion by 2022, retail is presumed to be one of the largest sectors contributing to itsgrowth. In the near future, we will use a combination of voice, chatbots, mobile and other channels to make shopping more effortless and more human. Brands will intelligently curate the individual customer’s experience in each of those channels with AI-driven insights. Consumers will be able to simply, verbally, ask for what they want and get it. As a result of all this, we may be able to make commerce a seamless part of the fabric of daily life.

Original article courtesy of

Amit Sharma

Original article can be found here