The promises of what artificial intelligence will be able to do in the near future are almost inconceivable: it’ll be able to solve climate change, take over some public policing functions, and solve worldwide unemployment. However, artificial intelligence is still in its infancy: imagine a toddler who’s mastered the very basics of walking and talking, but needs help feeding himself, dressing himself, and understanding right and wrong.

Artificial intelligence may sound like something that only the big players in eCommerce can benefit from, with the cost and data collection requirements making the likes of IBM Watson out of reach for most small businesses.

Yet, Gartner predicts that by 2019, “startups will overtake Amazon, Google, IBM and Microsoft in driving the artificial intelligence economy with disruptive business solutions” (content available for Gartner clients).

Some businesses, large and small, have already made the foray into using artificial intelligence in their eCommerce stores. Below, I take a look at some of the current uses of artificial intelligence, the software powering its use, and the future of artificial intelligence in eCommerce.

1. Visual search

“It’s grey, well greyish-green, and it has two straps, one of the straps is blue and it’s quite big but not too big and, well, let me show you a picture…”. Even the most eloquent of orators can find it difficult to describe a simple item to someone else. As the old adage goes, a picture is worth a thousand words.

I, for one, have spent many an evening scouring the internet for something I’ve seen and forgotten the name of, proceeding to try thousands of combinations of words and phrases, only for the internet to tell me “nope, that product doesn’t exist”.

Here’s the solution: eCommerce platforms have developed to the point where embedded visual search capabilities allow users to upload images to find the item or similar.

If you’re no good with words, visual search can help narrow down your search results by comparing image pixels with others that are similar.

Pinterest

Pinterest is an example of visual search done extremely well. Capitalizing on its transformation from photo sharing tool to eCommerce platform, the app has embedded simple visual search capability into its pinnable images. By assuming that users are ‘pinning’ images that contain what they’d like to buy, Pinterest allows users to click the magnifying glass on the top right-hand corner of each image, which takes them to a resizable crop-box, allowing users to select whichever item in the image they want to search for. This will then return links and pictures of the item and/or similar items.

Users can further filter the search results by topic or tag to refine the search. Pinterest Lens also allows users to use their devices’ cameras to search for similar items. Pinterest’s Lens functionality has even been built into Samsung’s new Bixby artificial intelligence assistant.

The tech behind visual search: Slyce

Slyce is a visual search and recognition tool that identifies products based on pictures taken by users, which then takes them to the relevant eCommerce product page. Slyce has partnered with brands such as Urban Outfitters, Neiman Marcus, and Tommy Hilfiger, who recently used the technology in a runway show.

Slyce created a customized ‘Runway Recognition’ app for Tommy Hilfiger with both 3D and 2D functionality. The app allows attendees to snap images of clothes live on the runway, or of event signage and advertisements, and then takes them to the relevant item page on Tommy Hilfiger’s eCommerce site.

 2. Conversational and visual commerce

We’ve talked a lot about chatbots and their uses (and limitations) in the past, but how about combining conversation and visuals? As with standard visual search, it’s easier to show someone a robot a picture of what you want than to describe it (who has time for charades, anyway?).

The tech behind blended conversational and visual commerce: Mode.ai

Mode.ai is an artificial intelligence-powered visual chatbot that functions through Facebook Messenger. Users upload pictures through the chat interface, and the bot will then search for the same item or similar items so that the user can buy the product online. The user can ‘talk’ to the bot and guide it through its search process with prompts, such as ‘no, that’s not right’ and other clarifying phrases.

Mode.ai works by trawling through millions of photographs from eCommerce retailers, and allows users to shop directly from stores.

 3. Virtual personal shopping

It’s likely that only a handful of us have ever used a human personal shopper, so what’s the appeal of a personal shopping robot? Can we trust a robot to know their MOM jeans from their mini skirts?

Generally, the higher your level of service, the more likely your customers are to stay loyal. The more you know about your customer, the more informed you are to make suggestions that align with their style, budget, and shopping habits. AI-powered personal shoppers can cut the small-talk, and take you and your customers from strangers to BFFs in no time.

Who’s using virtual personal shopping tools?

Shoptagr

Shoptagr is a ‘save it for later’ wish-list service, where users can browse over 1,300 retailers and save items they’re interested in buying. Shoptagr’s personal shopping assistant alerts users when products go on sale, are low in stock, or are back in stock.

Its predictive analytics bot analyses its customers’ behavior to learn their shopping habits, allowing the company to curate a personalized customer journey and influence users’ buying decisions. The online personal shopper, which works across devices, locates coupons for customers based on their history and advises them of the best location and time to buy an item.

Stitch Fix

Stitch Fix is another personal shopping service combined with an online subscription service. Customers receive five hand-picked items per month delivered to their door. They can keep what they like, and return the rest. Stitch Fix stylists – both human and machine – choose the best items for customers based on their recorded profile information such as budget, style, and lifestyle. Stitch Fix’s algorithms then match products to customers, and advise companies how much inventory they need to buy.

Stitch Fix’s algorithms can also:

  • Connect with customers’ Pinterest accounts to learn what styles they’re pinning to their own boards
  • Record customer decisions, such as whether to keep or return an item and the reason why, which powers item suggestions for their next subscription.

 4. Hyper-personalization

Ever felt like the websites you visit know what you want better than you do? As internet shoppers, or even just as internet users, we experience targeted ads all the livelong day, and isn’t it pretty cool that the websites we buy from are able to suggest items we may like on the basis of what we’ve bought previously? The truth is, this works: according to McKinsey, 35 percent of purchases on Amazon are made as a result of their product recommendation engine.

Software such as Nosto can retarget customers through email using data they glean from monitoring visitor mouse action, and provide personalized product recommendations across all devices.

However, retargeting customers once they’ve left your website is risky business, as even the most enthusiastic and seasoned of internet shoppers (myself included) can ignore your marketing efforts – nowadays, it’s going to take more than a lazy email addressing me by name and a product you think I’ll like to impress me.

The tech behind hyper-personalization: NeoWize

NeoWize takes personalizing the customer experience even further. Instead of relying on historical customer data which can help to focus retargeting marketing efforts, NeoWize collects real-time data while the customer browses the website.

NeoWize analyzes the customer’s shopping journey from the second they land on the website – which products they linger over, which products they scroll past and ignore, the clicks they make – in order to personalize the journey while the customer is still engaged and on-page. Extracting this real-time information increases the likelihood of conversion by presenting the customer with hyper-personalized suggestions. In short, it’s all about the here and now.

The future of artificial intelligence in eCommerce

The fact is, artificial intelligence as a whole is still highly limited. Google may now be able to conquer an ancient, complex, and highly intuitive board game, but the triumph is human. Without an ‘army of humans‘ recording millions of moves and adjusting algorithms, no machine could have beaten authentic human intuition and intent.

The same goes for artificial intelligence in eCommerce. It can recognize a keyword or pixels within a picture, but when it comes to understanding human intent or common sense, artificial intelligence falls short of the mark. It cannot, for example, cognitively understand why we make certain choices when shopping online. So, where does that leave the future of A.I. in eCommerce? We asked the experts.

Customer service

Chad Rubin, founder of Skubana, believes that the future role of artificial intelligence in eCommerce customer service will be of a much more remedial nature. He says, “eCommerce business owners need to be especially responsive when there’s a problem, to ease our customers’ concerns. Artificial intelligence could automatically diagnose an issue based on keywords in customers’ emails or chats, and send a standard “try this” response to help them troubleshoot their issues”.

Further personalization

Customers love personalization, and artificial intelligence is already making strides in enabling companies to lead consumers through a buying journey that suits them.

Mahi de Silva, co-founder and CEO of Botworx, says, “Gone will be the days of one-size-fits-all marketing; consumers will start to see highly customized marketing campaigns and incentives. These will be driven by machine learning and adapting to the real-time needs of a consumer, e.g. a weary traveler who’s missed a connecting flight and needs a last minute hotel room”.

Is your business already using artificial intelligence in your eCommerce store?

We’d love to hear about how businesses are currently using artificial intelligence to power their online stores. Let us know in the comments below about your experiences or email me at rhian@getapp.com.

To keep up to date with the latest trends and insights, sign up to our free email newsletter.