AI in E-commerce. Everyone keeps talking about it, how many know how to do it?


AI in E-commerce. Everyone keeps talking about it, how many know how to do it?

The use of artificial intelligence allows you to automate and streamline many different activities. AI makes fewer mistakes than a human and can handle more customers. The orders placed are processed faster by it, which means that the ordering party receives the desired shipment faster.

Although advances in artificial intelligence since the 1980s are far less spectacular than promised, any online store owner can now take advantage of the opportunities that it offers. Let’s see what benefits it will bring when used by our e-commerce.

Improving the accuracy of a store search engine

Many customers leave our online store website if the results returned by search engines do not meet their expectations. Nobody likes it when most of the phrases searched do not lead to finding the products or information we want.

An example might be Twiggle, which contextualizes search phrases, trying to understand their true meaning, and guessing the motives of the person trying to find a given phrase. The AI divides it into parts, for example, if a user enters a black table with three legs in the search engine, Twiggle recognizes the table for the search item category, and the “black” and “three legs” for the attributes of that category. Then AI searches in the product database all the items described in the “table” category, which have desirable attributes: “black”, “three legs”.

Twiggle’s way of thinking is similar to the way a person thinks, it understands the typed words in the search engine by understanding grammatical rules and relationships between words.

Twiggle is currently only available in English, and improving its code so that it understands flexo languages, for example the Polish language which seems to be something difficult to achieve in the near future. If your shop attracts many English-speaking customers, you may want to install the program yourself at your business.

If you are operating on the Polish market, you should familiarize yourself with the Unified Factory solution. Smart search, or “clever search engine”, can understand what we are really asking because each query is contemplated by context, not by individual keywords. The query converts to a mathematical string, so it is able to search available resources at a rapid rate and distinguish, for example, juice with carrots from carrot juice.

Another example might be Custom Training. It is able to learn to understand the content of images so that it can search for items in the image. If a customer finds that they liked a T-shirt that their friend was wearing, and they have a picture of this T-shirt, they can use Custom Training to search for the same model of the T-shirt.

If your store has a large number of products that attract customers mainly by the way the product looks, then Custom Training might be the solution for your e-commerce!

Intelligent chatbots will improve customer service

Instead of the client using the search engine, chatbots will make it possible for the customer to search with the help of AI to find the product or the get the information they need. Artificial Intelligence of this type is capable of conducting simple conversations that are essentially no different from normal conversations conducted by customer service. Based on the information provided by the customer, the chatbot can scan the product database for the item closest to the wishes of the person.

Chatbots can also be used as a supplement to standard customer service processes. Answers to many of the questions often asked by buyers can be provided by artificial intelligence, thanks to which employees can save time on simpler, more repetitive tasks and focus on more demanding cases.

Virtual Assistants

Similar to, but more advanced compared to chatbots, virtual assistants like Siri and Alexa operate. Alexa helps Amazon’s clients find interesting events. Thanks to her help you can order selected food products one day and have them delivered to your home the next morning.

Another skill that virtual assistants and more advanced chatbots possess, is the ability to answer the client’s very specific questions concerning a particular item. If the client wants to find if a given camera is waterproof then AI will use the complete database available to deliver a correct answer.

The purpose of having virtual assistants is to obliterate the difference between contacting a real retail worker working in a traditional, physical store and AI helping buyers shop online. The concept behind this solution is the idea that all the advantages of stationary stores should be replicated in e-commerce.

Creating your own virtual assistant is still too expensive for most e-commerce shop owners. Nevertheless, it is important to remember this option as technological changes may soon make it available to the average entrepreneur.

Better Methods for Analyzing Consumer Behavior

Consumers interact with your shop through a variety of different channels. This makes analyzing their behaviors and preferences much harder. Thanks to artificial intelligence, Boomtrain, it is possible to monitor the client’s behavior in each of them simultaneously. AI will analyze the interactions made on your e-commerce website and will see how much the latest marketing campaign was liked by the clients. It will also inform you when to send messages to clients, for example, information about product sales to maximize the number of transactions.

Make better suggestions for users

Efficient AI is able to improve its recommendations based on user account data. The history of a person’s purchases allows artificial intelligence to construct quite precise approximations of their preference. As a result, the proposed products are closer to the needs of the consumer, which increases the chance of the consumer making a purchase.

An example of the use of this method can be Steam, which, based on the analysis of computer games owned by the user and the groups / curators he belongs to, makes a recommendation for titles. The titles thus selected have a greater chance of being of interest to the consumer than if they were randomly selected. Netflix or Spotify use similar systems.

Personalizing a customer-led offer is a good idea for every online store. AI used for this purpose is simpler than others in this category, which should facilitate its implementation in our e-commerce.

Caring for the Quality of User Reviews

Scams in online stores are reviews written by users who do not own the product. Their authors may be fans or anti-fans of the brand, and their activities negatively affect the credibility of our site. Although the creation of good AI, which with great accuracy identifies such reviews is only in the plans, it is now possible to reduce the number of false opinions.

Amazon uses artificial intelligence that weighs user reviews based on several factors. The opinions of people who have made a verified purchase of an item or received a lot of positive reviews from others for their review are promoted. Over time, the system will be improved, which will increase confidence in the ratings of users left behind with the products.

Over the past few decades, AI has made very slow progress. Nevertheless, it is becoming increasingly widespread in everyday life. Even if the artificial intelligence does not fulfill the expectations that are assumed and its function more closely resembles more or less simple scripts than the Star Wars robots, it already benefits those who can use it.

Artificial Intelligence in e-commerce is not a song of the future, it is happening now. Often we do not realize how many items are already automatized in order to help us shop faster or to find an answer on our own. When thinking about automation in the context of e-commerce, unattended logistics and speeding up customer service comes to mind, and yet Alibaba has launched a fully-fledged chatroom with customers in 2015, and two years later (2017) also provided a chatbot to help salespeople operate numerous transactional inquiries from their customers to the mentioned Twiggle. Algorithms help us to do more shopping, they suggest what we should watch, or what music to listen to.

Just a few years ago in the tourism industry it was assumed and expected that robots will first help clean the rooms and receive guests at the reception. Yet time has shown that artificial intelligence has completely automated the process of buying a trip, booking a hotel, or purchasing an air ticket. In this all we are helping chatbots, which are becoming more and more intelligent and increasingly popular.

Companies such as Skyscanner, Expedia, and have already dedicated teams to develop new technologies. When nearly 60% of text searches (chat, email) in a company are a repeated question, why not use tools like smart search, smart FAQ or just chatbot that will automatically deal with them? The customer will be happy because he will get a quick response, and our service team will get more time on more complex issues.


Marek Zmysłowski

I build companies outside the First World, so it becomes First.

Sign up for my newsletter

Where I share what I learn in Business, Tech, Startups, and Longevity.

and get 100% discount for my ebook