Commerce is no longer associated with physically buying and selling items. Yes, many of us do enjoy the experience of going into a store to choose a new piece of clothing or a pair of shoes. But when it’s all about convenience and comfort, people go online. Not to mention that it’s also cheaper.
And if you’re lucky enough to live nearby an Amazon operational warehouse, you could probably get the item in a matter of hours. Scratch that. How about in minutes and during a holiday season? That’s the power of eCommerce that we’re going to get used to soon. With a progression like that, ‘same-hour delivery’ will soon be a standard for online shopping. But it’s not just about the convenient logistics of online shopping. Modern online shopping offers unparalleled flexibility of access and selection of available interfaces. You can shop away sitting in a bus. You can buy your next favorite gadget while relaxing on a lawn on a beautiful sunny day. All thanks to the ubiquity of smartphones.
The graph above represents important regional economies. We’re not even talking about countries like India, where the adoption rate is about 20% but the actual number of people behind the percentage amounts to hundreds of millions. Most of these people will try to shop online using their smartphone.
That’s why mobile commerce is such a fast-growing niche that caters specifically to smartphone users and their respective app ecosystems. Mobile commerce has its own ‘rules of the game’, either dictated by software or hardware limitations. And retailers are starting to go for custom mobile app development to adjust to them. AI is taking eCommerce by storm, and smartphone users are being slowly engulfed by it.
Let’s take a closer look at some of the most promising applications of AI and machine learning in mobile commerce to find out how businesses can leverage them today. Mobile traffic just recently surpassed desktop visits, so these applications are potentially useful for any online business as the number of smartphone visitors will keep growing even if mobile users are not your primary audience at the time.
Search and Navigation
You have to bear in mind that a smartphone has limited real estate, and often this space is taken up by images. That’s especially true for eCommerce development where product images have a major role in web design and, thus space allocation. This fact could lead to some poor design choices that will hinder a user experience and site navigation.
And if a user is not able to navigate to the page they want to, they’re going to turn to site search. So you either have to: a) make changes to your mobile design to facilitate smartphone users, or b) refine and customize your site search so it could serve as a substitution for your navigation.
The beauty of the second option is the fact that you will also improve the search experience for desktop users. So this should be a natural choice. Given that site search can be a game-changer for your conversion rates, its optimization should be treated seriously by both mobile-first and desktop-first eCommerce businesses.
Luckily, companies like Twiggle offer AI-powered eCommerce search that’s specifically built to accommodate shopping searches and provide a richer product context depending on search queries. It is being backed by big names, like Alibaba, so you know that they mean business.
Swiftype is a similar product endorsed by big commerce names, like Shopify. But it also offers enterprise search solutions, so companies that have a POS system and offline venues could improve their UX as well.
But we can go deeper. Search boxes might also be a thing of the past pretty soon.
Some of the core audiences for mCommerce are already very comfortable with voice-activated search. That’s why companies like Voysis are trying to enter the market with unique products geared towards mobile shoppers. They even promise a 10x reduction in ‘homepage to cart’ time, which is a pretty bold statement. But these guys are starting up. Companies like Celebros offer ready-made site search solutions that are powered by machine learning (NLP) to deliver natural language search that gets better as the more people use the search function.
This mCommerce niche is the most promising and comfortable for AI. All of us have at least one messaging app on our smartphones.
That’s over 3.5 billion people actively using messaging apps to communicate. With these numbers messaging applications can be considered a vital, universally available and adopted form of communication.
AI fits perfectly in this ecosystem because it can easily substitute a lot of the activities previously performed by humans. The growth of chatbots and digital assistants is a testament to this fact. These solutions are starting to successfully replace customer care representatives and even sales managers by doing a good job of informing the customer and even helping them make purchases over their smartphones. Chatbots free up company resources by adding a layer of automation to various internal and customer-facing processes.
What’s important at the implementation stage is the understanding of the fact that there are different kinds of AI-powered solutions and those that mimic AI. For example, there are AI-based interfaces that have a profound understanding of the particular task that you want to accomplish with their help. The most notable and probably familiar example is the Clippy – Microsoft Word assistant that many of us had a chance to use. There’s not much interaction that resembles messaging via an app.
Then there are rule-based chatbots – bots that can respond to specific pre-determined questions or act based on your input, but only if that input is recognized. The discussion on whether these are a match to a full-blown AI-powered chatbot is still going. Right now the situation with many of these solutions can be illustrated with the following picture:
The top tier that involves artificial intelligence is AI/machine learning powered chatbots. The defining feature of these solutions is their grasp of NLP or natural language processing. This machine learning niche deals with semantics and context of human conversations. It’s not about recognizing words, but about identifying the context of the discussion to participate in it. For example, an AI-powered chatbot will understand that when you’re saying that you ‘bombed last night’ it means that you poorly performed something and not that you were involved in an actual bombing. There’s even a special award called Loebner Prize, which is a Turing test for AI that’s often used as an evaluation tool for chatbots. It determines how ‘human-like’ the conversations are. Many of the recent winners are chatbot solutions geared towards commerce. Let’s take a look at some of the notable AI-based chatbots that were designed specifically with eCommerce in mind.
BotCommerce is a chatbot designed to work specifically with Magento and Shopify stores. So it covers the majority of SMB eCommerce stores. Its features include customer matching (engage people with the bot, if you have their phone number) and order tracking commands to cover the basic needs of a small online store. Mr.Chatbot is a similar solution with an emphasis on the ability to engage in conversations and process natural language requests more fluently. At least, that’s their stated forte.
LetsClap offers a CRM system for customer care and sales, which also has its chatbot product that integrates with the rest of the system. So you can potentially see chats performed by bots and reassign conversations. Morph.ai is another eCommerce focused chatbot that is more integrated and can be added to your store in one click. Supported platforms include Magento, Shopify, Woocommerce, OpenCart, etc. Clare.ai is a good example of a specialized bot, although it’s focused not on eCommerce but on banking (we wanted to include it to demonstrate the additional functionality that a specialized chatbot could offer, once it’s programmed with specific tasks in mind). Msg.ai is a robust platform for conversational commerce that can reach your customers through practically any messaging solution, and is fully integrated with some of the most adopted sales management systems like ZenDesk and Salesforce.
AI for Video
Video consumption is one of the core functions that people use their smartphones for, just like messaging. That’s why Facebook Messenger is the second most used app on smartphones, followed by Youtube.
Videos are easy to digest via smartphones, so they offer some unique marketing opportunities for the mCommerce scene. And AI is here to help.
There aren’t that many direct applications of AI for video content. If we’re talking about the medium’s complexity, then videos are probably at the top of the list. They’re dynamic and require much more processing power, even for a human. So machines also struggle with actual video analysis that would work in a meaningful way for mCommerce marketers. That’s why language and image processing AI solutions are far more abundant these days. Images and text are static and more predictable in terms of programming AI’s understanding of them.
However, there are companies that are trying to break that barrier. Affectiva is a startup that recognizes human faces and tries to interpret their mood via a video link. So, presumably, you can have someone watch your product video and use Affectiva to analyze the emotions that a person was experiencing while watching it. It uses smartphone sensors and camera to identify these patterns. The premise is a bit creepy, as you’ll essentially have a machine watching you. Sort of like Hal 9000 from 2001: Space Odyssey. Due to the complexity of the technology, it’s also expensive at this point, with prices starting at $1 per one minute of video analyzed. So if you’re a small eCommerce business with hundreds or even thousands of product videos, this might be too heavy a burden to carry.
What AI is good at is recognizing customer-centric patterns related to video marketing and interpreting these results for you.
Magisto uses the power of AI to customize video content based on the video-watching audience. IRIS.tv exploits AI to continuously optimize video assets to better suit specific audiences and niches. It’s similar to Magisto but comes with its own BI tool that potentially can help you extract even more insights out of the video consumption patterns.
There are tons of potential applications of AI in mCommerce. Some of them are limited by the technology, like with video content. And some haven’t yet been implemented or even envisioned. It’s important to understand that the line between eCommerce and mCommerce is practically gone. Long gone are the days where you’d debate whether you need a separate version of your website or a responsive re-design to match the demand. Google made it clear a year ago that you should have a website that works fluidly with desktop and mobile users at the same time. In this ecosystem, mCommerce and eCommerce merge to become commerce.
That’s why AI applications will eventually lose their specialization or its remaining features. So many chatbot apps that are being directly advertised as compatible with Facebook Messenger, which means with mobile users, will eventually drift towards the famous Clippy assistant that we mentioned. Voice commands will become more common as users will get even more used to Siri, Alexa and other digital assistants. Nobody would want to go into a chat window and type something in. It’s going to be easier just to ‘ask your website’ for help.
Seamless experiences that work across all platforms and a variety of mediums are going to be a standard for online shopping, powered by the ever-growing potential of AI and cloud computing. Six years ago Gartner predicted that by 2020 most of the consumer interactions online would be handled without any human input. As you can see, we’re well on the way there with the ‘explosion’ of AI applications that augment online shopping and selling experiences and transform the velocity of commerce.