
Every purchasing process is preceded by a complex decision-making process. For the human brain, this entails a long struggle involving an emotional impulse, a wish and the weighing up of facts and prior experiences. But one thing is certain: the customer feels happy when he or she finally buys something. This means that a good prerequisite for the consumer society would not be an endless choice and variety of consumer goods, because too many options simply overwhelm the human brain. Instead of being happy about a selected product, purchasers feel frustrated about all the products they have not chosen and dissatisfied with what they have actually bought.
Many people decide against buying when there is a slew of choices and end up leaving the store empty-handed. So how can a retailer provide shoppers with the perfect selection for each individual customer? This is exactly where artificial intelligence – AI for short – comes in with a wealth of possible solutions. But beware! Not everything that is called AI is actually AI.
AI is not just about the ability of an IT system to behave in a human-like and intelligent way. Rather, it is all about imitating other core abilities such as perception, understanding, action and learning. Artificial intelligence can be seen, for example, in the areas of machine learning and deep learning.
Machine learning describes a process in which computer algorithms learn from data to recognise patterns and correlations or to show desired behaviour. There is no need to program each individual case explicitly. For example, algorithms in online stationery shops learn that there are certain classes of writing materials that certain classes of customers buy. This happens without any prior definition of what calligraphy pens are or what creative hobbies young mothers are keen on.
Another method can also be used for image data labelling and annotation. In this process, people label pictures with information on whether a face appears happy or sad. After several thousand or ten thousand examples, an algorithm masters the task of classifying new images itself. This application of machine learning is an enhancement of algorithms and just one of the many tools of AI. Machine learning with large artificial neural networks is called deep learning.
In the retail trade, the main areas of application are found in marketing, e-commerce, logistics and customer experience management, as well as in the optimisation of warehouse and delivery capacities and in intent recognition, in which customers' purchasing intentions are recognised early on.
In cooperation with the EECC (European EPC Competence Centre), which is heavily involved with the topic of block chain and artificial intelligence, the German retail giant Real has recently tested a personalised coupon platform known as predictive commerce. In this process, customer and product data were correlated to predict which coupons would be redeemed and when. While the redemption rate for classic coupons is usually 6 per cent, these coupons were issued in real time and 57 per cent were redeemed. This clearly showed that AI optimisation – as opposed to conventional methodology – led to a better result.
A study conducted by EHI (Acar, Spaan & Gerling, 2019) also showed that AI-based technologies will significantly change the retail industry in the coming years: areas where there is currently a shortage of skilled workers will be automated, trend forecasts will be revolutionised, new jobs will be created, and solutions for ecological challenges generated.
Many retailers do not yet see the opportunities that AI solutions offer, but they are the future. This is because AI helps retailers to analyse customer needs and desires and so optimise processes to ultimately achieve better results.
Introducing AI into companies requires the cooperation of the department concerned with IT experts
External service providers can provide support with already existing applications from other industries
Retailers should follow a long-term strategy with small pilot projects to gain experience and then design their solution
Cloud-based platform solutions have a high degree of maturity in the application of master data (e.g. for validations and AI-supported classifications) and are thus a good starting point for AI technologies in the retail trade
About the author
Stefanie Otto is a junior project manager specialising in retailing and retail technology with gmvteam GmbH, a Düsseldorf-based innovation agency for retailing and urban development. She is also an author on the Zukunft des Einkaufens blog.