Dusseldorf The small cameras hang everywhere in the Edeka Center in Warnow Park in Rostock. Your task is to constantly take photos of the goods shelf opposite. With the help of artificial intelligence, software translates these images into structured data. The cameras cost less than 200 euros each.
But they have a big impact. With this, managing director Stephan Cunäus knows at all times where there are gaps on the shelves, which products are sold out and which are in the wrong place. Since his employees can now fill the shelves more quickly and in a more targeted manner, the availability of goods has increased by three percent after just a short time.
This effect can also be observed at the French retailer Monoprix. Chief Operating Officer Damien Pichot has already converted it into sales: After the technology company Captana introduced technology in 100 of its stores, the branches have had an average of 1.9 percent more sales since 2021.
At the same time, the costs at Monoprix have fallen. Since employees no longer have to constantly walk through the stores to check the stock, Pichot saves around 400 working hours per store per year. The investment for the 120 cameras per store pays for itself very quickly.
The examples show that artificial intelligence (AI) is no longer just a topic of the future in retail. In many areas, AI is already in daily use. Cameras on the shelves and smart forecasting models organize replenishment, algorithms set the prices, and self-learning systems control the deployment of staff.
“Artificial intelligence is a very big opportunity for retail,” confirms Matthias Guffler, partner at the strategy consultancy EY Parthenon. “It allows him to react better to many challenges.”
And the technologies bring tangible benefits to retailers. “Overall, based on our projects, it can be assumed that costs of between 15 and 30 percent can currently be saved, depending on the area of application,” says Guffler.
“For some applications it is already common practice in many companies, for others the first tests are now starting and are bringing surprisingly good results,” he observes. AI can help in many areas, from disrupting the supply chain and digitizing processes to personalizing the customer relationship.
Image recognition in particular can be a huge driver of efficiency in retail. “It can, for example, monitor the presentation of the goods and thus save a lot of work,” says the EY Parthenon expert. “The faster detection of gaps in the shelf can directly increase sales,” says Guffler. For retailers with several thousand stores, that can be a pretty big lever. And the customers also benefit from a better offer.
Artificial intelligence: German retailers are pioneers when it comes to technology
In addition: “The technologies are becoming cheaper and more profitable as a result,” says Michael Unmüßig, CEO of Captana, a subsidiary of the French tech group SES Imagotag. The company develops cloud-based sensors based on electronic price tags, including the cameras used at Edeka and Monoprix. All applications have been checked in accordance with data protection law, assures Unmüßig.
You can see the effects in the profit statement within a few days. Michael Unmüßig, CEO of Captana
“You can see the effects in the profit statement within a few days,” says the Captana boss. He sees German retailers in a pioneering role: “They have recognized the potential of AI,” he says.
Dusan Rnic, European boss of the American start-up Symphony AI, has had similar experiences. “It is currently very easy to convince German retail managers to use AI solutions,” he reports. “You can see immediately how quickly it pays off.”
Symphony AI develops software for the use of AI in retail, from supply chain optimization to on-shelf image recognition, and works for dozens of European retailers. The drugstore chain Rossmann, for example, uses software from Symphony AI to optimize replenishment planning for each individual branch.
The openness of German retail for AI confirms a survey by the retail research institute EHI among 92 large retail chains, which together represent 60,000 branches in German-speaking countries. 69 percent of them are already using AI, and another nine percent are planning to do so in the next two years.
The main areas of application are the forecasting of demand and the planning of replenishment. More than every fifth retail chain sees advantages of AI in pricing. 17 percent of retailers mainly use AI in dialogue with customers. However, this area is likely to become even more important in the future with the rapid development of chatbots such as ChatGPT.
Metro and Douglas increase sales with artificial intelligence
A pioneer in the use of AI is the wholesaler Metro. Around 60 employees work there full-time on such applications. The company has already invested around half a billion euros in the technology in recent years. “The use of artificial intelligence measurably increases our sales,” says Timo Salzsieder, IT manager at Metro.
Metro is already using AI in numerous business areas. A big advantage: unlike the retailers, the wholesaler knows every single customer because they identify themselves at the entrance with a customer card or app. “That’s why we can give each customer individual product recommendations with the support of AI,” says IT boss Salzsieder. This also has a direct impact on sales.
>> Read also: Obi boss Sebastian Gundel relies on digitization
The drugstore chain Douglas has been testing AI in forecasting demand for almost a year. Machine learning, for example, is used here, in which algorithms can recognize patterns in data sets and develop solutions from them. Konstanze Gallinatus, who is responsible for the area at the company, says: “We are already seeing clearly positive effects.”
Demand at Douglas has always been characterized by seasonal effects. Now there are strong trends, for example when influencers promote certain articles. “Corona and now inflation have made the forecasts even more difficult because the old patterns no longer apply,” reports the Douglas manager.
As a result, the challenge of precisely predicting the demand for individual products is increasing. “Machine learning helps us a lot because it quickly recognizes signals in the data and learns from them,” says Gallinatus. Any number of variables can also be included in the forecast, up to local holidays or the weather forecast.
Artificial intelligence will not replace humans
But the manager also sees the limits that the use of AI still has. “Artificial intelligence is only as intelligent as the database allows,” she explains. “The more data you have and the more frequently the customer buys from you, the more it helps you.” At Douglas, where many customers only order a few times a year, it is therefore difficult to use AI to personalize the customer approach to use.
>> Read also: Will artificial intelligence help employees – or replace them?
The interaction between humans and algorithms is also very important, adds Gallinatus: “Artificial intelligence learns independently. However, through targeted training by people and their experience, it provides us with significantly better results.”
Trade expert Guffler is also certain that AI cannot replace humans. But he changes the requirement profiles for the staff. Employees would have to learn to work with the AI and be able to use it optimally. He emphasizes: “Artificial intelligence is a tool for making better decisions, not for making decisions for people entirely.”
More: These six industries could be revolutionized by artificial intelligence