BearingPoint as the partner of the ESCP “Retailing 4.0" Chair is committed to contributing to building bridges between the academic world and companies. As part of this engagement, the BearingPoint Retail, Luxury and CPG team mentored the writing of a series of articles of ESCP MSc students and shared insights on major retail stakes. Enjoy your reading !
We live in a world where Artificial Intelligence (AI) and Machine Learning (ML) are slowly invading our everyday tasks as consumers, suppliers and sellers. China is currently the world's second largest AI ecosystem, just after the USA. The Asian power is investing billions of dollars towards the development of AI in its country. The Ministry of Science and Technology of the People’s Republic of China appointed the 4 largest Chinese AI companies: Baidu, Alibaba, Tencent, and iFlyTek to take the onus of the development of AI throughout the nation across several dimensions.
AI and ML are completely revolutionizing economies, ensuring increased efficiency in production, improving profits and reducing costs. It is helping brands making better predictions and consequently more informed decisions. But what are ML and AI concretely?
Machine learning is the “field of study that gives computers the ability to learn without being explicitly programmed” (Arthur Samuel; 1959). Machine learning involves application of computational methods that uses experience to improve performance and make accurate predictions. Machine learning requires loads of data about the customer in order to function effectively.
"Artificial Intelligence is a branch of computer science concerned with the study and creation of computer systems that exhibit some form of intelligence: systems that learn new concepts and tasks, systems that can reason and draw useful conclusions about the world around us, systems having the capability to assume the meaning of the natural language or perceive and understand a scene, and systems that perform other types of feats that requires human types of intelligence."
Simultaneously, since AI and ML require tonnes of data to function effectively, they are paving the way to an increase of ethical concerns, ranging from how reliable AI systems can be, to violating human rights and values such as privacy.
Therefore, in this article, we will try to answer the following questions: Do the benefits of AI and ML for retailers counterbalance their disadvantages concerning data protection?
The first point of this paper will be about the various benefits AI and ML bring to the customers, as they allow more personalized products and services, thanks to new technologies and devices. Those benefits can be reached online (web) and offline (stores), thus enriching the omnichannel customer journey – their key driver for companies being the ability to collect (personal) data on customers, thus letting them invest a lot towards data collection and science.
However, these facts raise questions, which will constitute the second part of this paper: are people sure that those data are used in a proper way? Are people sure that they are not used for other purposes (e.g. monetization)? Are people fully aware of what usage companies do with the data? Are people really benefiting from these deals with companies: “data vs. better services”? Facing the increasing power of data/tech companies (e.g. Google, Amazon, Alibaba, etc.), these questions raise the need for a counter-power - not letting its monopoly to those companies, in order to protect people and to avoid a misuse of their data. This is clearly what multiple governments are trying to achieve within their countries/regions although tech companies are moving forward faster than governments. This is becoming increasingly critical as the pace of innovations is clearly booming.
Therefore – the last point of this article – a good move for companies should be to anticipate those concerns by being transparent as much as possible to gain the customers’ trust, thus more willing to purchase. Last but not least, providing a relevant legislation in order to protect people should be an initiative taken by governments to avoid that companies monopolize power.
Most of the examples we are going to mention come from China, which is also one of the most technologically advanced countries when it comes to AI and ML. Moreover, in general the Chinese population is also willing to share data with firms. In fact, customers are eager to adopt new technologies, as they are convinced that their data will be used only within the country. Trust related to the use of data is therefore higher in China than in the U.S. or Europe.
Retailers today are benefiting from Machine Learning and Artificial Intelligence from supply chain to customers’ experience, and this tendency is here to rise. Looking at the graph below, we can see some of the main benefits at a corporate level.
But how are retailers actually using ML and AI to their advantage when it comes to customers?
First of all, the key for a successful sales and marketing strategy for every business is to know their customers and this is exactly what retailers are doing: they leverage ML and AI to a better understanding of their clients. For instance, we can refer to Alibaba – the Chinese tech giant. They use an AI-powered algorithm called Tmall Smart Selection, that is “backed by deep learning and natural language processing” to help shoppers by recommending them adequate products, which will then communicate to the retailers in order to increase inventories and be able to keep up with the demand. This way, it is easier to optimize customer satisfaction, as retailers know what clients want. It is also simpler to meet customers’ demand and, in some cases, even anticipating their requests or needs before they even know they want something. It is a simple “win-win” situation.
Moreover, when customers feel “known” and “understood”, they are more likely to buy additional products, as the retailer can show a personalized selection of products after putting goods to their basket, inciting them to consume more. This also decreases the chances of leaving for a competitor, thus retaining the clients and increasing sales.
With this look-a-like product recommendations, brands can prospect, retarget, cross-sell and up-sell. These recommendations are based on the purchases previously made by the user or even those purchases made by other users who bought similar items.
Furthermore, there is a set of AI techniques that seeks active responses from users to comprehend their personality and consequently find products or services that match with it, encouraging the customer to purchase. It is indeed possible to use data for better consumer experience, through a gamification of the feedback mechanism. An example of that is Stitch Fix, which incentivizes customers to give feedback and turns it into a fun experience in which they swipe right on the items they like. Similarly, Feedier is gamifying surveys by providing customers with interactive forms and brands with real-time data on satisfactions. Customers get a fun experience and at the same time retailers collect data on what customers like.
Retailers are also improving their services for customers to the next level by having highly efficient chatbots. Those chatbots can understand nearly 90% of human queries. The newest ones can even depict human emotions in the messages and thus redirecting the customer to a specific manager to deal with their issues, warning the latter about the client’s emotional state. Not only do chatbots improve efficiency, as retailers need less employees, but it also enables the storing of data about each person answering the chatbot, thus anticipating future trends that may arise amongst groups of customers.
Another way retailers are leveraging ML to the benefits of their customers is with pricing. Willingness to pay can be measured by a person’s behaviour online, such as the products they purchase, and the time spent on a certain page. These pricing algorithms can analyse the prices of competitors for similar products in real time. Moreover, by simply analyzing the client’s recent purchase history as well as other external factors (e.g. seasonality, external events related to them, etc) they can and adjust the prices. This optimization in the pricing strategy is key to boost sales online.
Additionally, in the past years, customers have been increasingly willing to get greater value, convenience, relevance, status, authenticity and social connection. Even if many have privacy concerns, it seems clear that most of them want to be recognized - at least when it gives them some sort of added value (e.g. discounts, lower prices, personalized recommendations). For instance, according to Trend Watching around 54% of customers expect to receive a personalized discount when first getting to know a brand online, while a 71% gets frustrated in case they are offered an impersonal shopping experience.
Furthermore, the way customers research on Google has also changed in the past years, which justifies the 60% increase in the “... for me” that has been registered.
In a nutshell, AI and ML are helping deliver enhanced customer experience, meaning more personalized and convenient interactions. They do not only translate to better results across KPIs (e.g. for increased propensity to buy or higher average order values) but also give brands the ability to provide a more personalized experience, a better assistance to the consumer and solve the problems of clients – who end up overwhelmed due to the vast online choice. AI and ML, however, are not just impacting our shopping online. In the next subpart, we are going to see in practice how AI-empowered tools are impacting the customer experience in-store.
AI enables us to anticipate customer expectations: it uses the customers’ search data or browsing history, social network interactions, physical store visits, etc, to anticipate their needs. Retailers can track what is in stock and create better product collections to tailor efficient shopping experiences. A proof of how that can work is the Amazon’s 4-star retail stores in which products that have received a multitude of 4-star ratings online, are offered offline as well.
Furthermore, AI makes the shopping more convenient and improves the in-store experiences thanks to new technologies such as robots, beacons, smart mirrors and vending machines.
Robots shopping assistants help navigate customers through the products, identify items and eventually start conferences with a human assistant. The robot can be a trolley and shopping assistant simultaneously and arrange the home delivery of the purchased items in-store. Moreover, it is used to provide better and more sophisticated recommendations, as it learns from shoppers’ behaviour. These advanced machines are also used by retailers to assess the stock level on shelves and detect price errors.
Moving on to the next technology revolutionizing physical stores: beacons.
Beacons are small wireless devices that transmit a continuous radio signal to phones. The ID is then sent to a cloud server that consequently pushes content targeted to the owner of that device. To make a practical example, when a potential customer walks past a store while this store is having a sale, then the person receives a notification, inviting him/her to enter.
The third revolution are smart mirrors, which allow people to virtually try on clothes - that are in stock or not - in-store. They can then place an order from the mirror and have the clothes delivered at home. Such technology enhances the customer experience, helps increase sales and escape from the buy-try-return cycle.
Vending machines will also revolutionize the way customers purchase, even in industries where no one would expect to see them. In car showrooms, customers combine features (like colour of interiors, power of the engine), can pick the car they want to try and later talk to the salespeople.
These revolutions aren’t only meant to be used separately. Vending machines and smart mirrors together, for instance, will innovate the way customers purchase makeup items. With AI technologies, clients can see how specific eyeliners or lip glosses look on their skin and have the ability to make the purchase directly from the machine.
AI helps brands make the in-store experience more convenient also in terms of payments. In fact, it enables smart checkouts. Companies like SmartCart have invented shopping carts with cameras that can automatically add up the items the user is purchasing, prepare an order, and enable payments through a mobile device.
Similarly, AmazonGo offers the “walk-out-the-store” concept – the order will be charged to the individual’s Amazon account directly, enabling the customer to skip the “cashier step”, saving him time.
Another important technology, regarding smart checkouts, is facial recognition. It isn’t only meant to be used as a way of payment – customers pay with their faces – but also as an emotion detector. Indeed, facial recognition systems have a software that is capable of reading human emotions, thus identifying when a customer gets frustrated and warning the staff, so that the situation can be handled accordingly.
The challenge brands has to face is getting more personal and better at knowing their customers, in order to boost their performance. Additionally to the software in facial recognition, they can use emotional data and eye tracking. Walmart is already experiencing this and has a patent that recognizes the emotional state of customers. Knorr is also offering hyper personalization by offering meals depending on a customer’s Instagram feed. In San Paulo’s metro in Brazil, billboards recognize commuters’ emotions and display the “right” content; they base assumptions on how crowded trains are, what time of the day it is and the facial expression of people.
Similarly, AI and ML enable brands to become better at recommending products and driving conversions through learning – just like we’ve seen earlier for the benefits online, at the beginning of this part. An algorithm assesses all consumer responses to a series of simple questions and then leads to the recommendation. It determines how successful each recommendation is at converting consumers and practically learns from historic data which combination of touchpoints and recommendations are more effective, adjusting it consequently.
As we have seen, the future is the digitalization of all retailing channels. AI has revolutionized the customer journey and this is what a grocery shopping experience looks like when such new technologies are combined to the offline. Retailers are now able to recognize a customer when entering the store – thanks to Wi-Fi or facial recognition. Once the customer logs in into the store’s app, he is able to see his shopping list, with the shelves on which the products are located, illuminated on his screen, to highlight distinctly their position in the store. Moreover, the smart shelves will indicate complementary products or special offers. The user can scan the products with a QR code or on the retailer’s app; this way the clients have access to information such as origin, carbon footprint or nutrition information. When exiting a store, the need to go through the check-out phase disappears; in fact, the items in the basket are visualised thanks to the RFID scanners and machine vision system. By having the personal credit card already registered in the system, the client is able to simply walk out the store as he/she is debited for the items directly on his/her card.
These innovations do not only concern large grocery stores, as even convenience stores in cities are getting more digitised. Shopkeepers will use a mobile app with custom analytics that shows what items are most likely to be purchased, thus allowing them to order when stocks are low.
All these benefits, online as well as in stores, from AI and ML almost seem surreal in their convenience. However, just like signing a pact with the devil, every joy comes at a price – the price being the citizens’ personal data.