Tilde and StrongPoint expert talk: AI will significantly change the way we shop 

How will artificial intelligence transform brick-and-mortar shops in the future? To what extent is AI being employed in commerce at the moment? Giedrius Karauskas, head of the technology division of the language technology company Tilde, and Rimantas Mažulis, head of the retail solution company StrongPoint for the Baltic states and Finland, discuss the implementation of AI solutions in retail business. 

GIEDRIUS KARAUSKAS, Tilde: As experts in the retail business, it would be interesting to talk to you about your clients, retailers who use your solutions, and discuss how AI is already performing or will perform in retail, but before talking about this sector, I’d like to ask how you perceive AI yourself? Do you apply it and implement it in your own company or, on the contrary, have a sceptical attitude towards it? 

RIMANTAS MAŽULIS, StrongPoint: Currently, we are mainly engaged in employee education; we have issued guidelines and recommendations, we’re raising our employees’ awareness of risks that come with the usage of, e.g., ChatGPT, and what actions may lead to data transfer to third parties. Currently, our company strictly uses AI on a personal basis; our employees use it to search for information that could lead to bigger things. We barely apply it to our processes: we somewhat use it in marketing, text correction, brainstorming, and searching for how AI could facilitate the development of our systems. In the near future, we intend to find experts who can tell and advise us on how we can practically apply AI in our business.  

GIEDRIUS: Let’s go through different parts of retail and consider how AI can change the retail business in the future. E-commerce is already doing this and is being widely discussed, but what will happen with in-store shopping, e.g., checkout: how can we make use of AI, what should be done, and how should we change the client experience and management of such business? 

RIMANTAS: AI is going to make a big change in physical store sales; it’s even changing as we speak. For example, there is so-called computer vision that works with AI, which helps to identify a product in a self-service checkout, making shopping easier for the buyer. You can bring some tomatoes to the camera, and it will capture and identify them without the need to search through a long list of products. This system is very precise because it can also identify products through, for example, a plastic bag, it understands the differences between the different types of apples, and it’s being gradually implemented at points of sale.  

Moreover, this solution can also be applied to the prevention of theft. Do you know that the goods most commonly sold in self-service checkouts are vegetables such as carrots or onions (smiling)? Why? Because it’s the cheapest commodity. People put bananas on the scales and weigh them as onions to pay less. And that’s only the basics because AI computer vision can not only identify what goods are being purchased but also detect other forms of fraud, as well as suspicious behaviour in the checkout area. The most important thing is that these AI apps are genuinely well thought out; everyone understands that AI shouldn’t make buyers feel uncomfortable, so even if it notices strange behaviour in the self-service checkout, this AI solution always politely indicates the product that the buyer has selected, suggests checking whether the choice was correct, and then, if necessary, sends a notice to the employee so that he or she can check the situation personally.  

There is another AI solution that we have already implemented in Estonia, and we are working on it in Lithuania. I’m talking about automatic age verification: AI scans a person’s face without capturing a picture and uses some algorithms to assess whether a person is fit to buy certain items with an age limitation.  

And, of course, AI plays an important role in autonomous shops where there are no cashiers at all. There are some of these shops in Vilnius, too. They have AI and computer vision cameras that identify the customer, as well as track and identify what goods the customer has picked up, and you only need to tap a card and pay upon reaching the check-out; in some cases, there is even no need to do this because the money is automatically retrieved from the customer’s account. 

GIEDRIUS: The implementation of AI solutions is largely limited by data protection. Even in our area, when working with chatbots, customers sometimes want us to train them with the data they have, but they can’t share it. What about retail, how is data protection viewed, for example, when it comes to automatic age recognition? 

RIMANTAS: The subject of data collection is always sensitive, but for this solution in particular, there is no such problem as it does not collect data. It doesn’t take any pictures that can be stored, it just scans certain features of the human face and determines whether a person is older than, for example, 25 years of age. If the customer’s attributes do not exceed the expected age limit, the program calls a member of staff for physical confirmation.  

GIEDRIUS: Retail businesses today usually have apps, especially in the case of large retail networks. They are used and loved by customers. How do you think AI could allow one to get even more out of them? Perhaps, for instance, by using existing data, AI could instantly draw up lists of goods and act as an assistant during shopping who answers questions about the composition of products, allergies, and where the goods can be found in the store?  

RIMANTAS: Well, it’s not even the future, it’s the present; retailers are already developing such solutions. For example, you can find a recipe on the app, and it informs you what you can buy for the recipe from that store. Moreover, it indicates where the goods are located in the shop, and the electronic label next to them may blink to make it easier to find the goods because it has geolocation.  

GIEDRIUS: Let’s change the topic: competition and a competitive environment. In such mature sectors, there are a large number of players, and competition is intense. If one player tries something innovative, the others keep pace. What about retail: do market participants believe that AI can make significant changes to the rules of the competition or monitor each other and follow each other’s footsteps? 

RIMANTAS: They are watching for a fact, and quite closely, especially in the Baltic market, which is not very large; everyone here knows everything about everyone.  I believe that retailers can employ many AI solutions in their internal processes, such as automated orders and sales data, where everything adds up automatically so that the store has enough goods for the next day. But it’s hard to find out about them because retailers don’t talk about them out loud. In most cases, when a certain company sets out to test a solution, the others monitor whether it has a significant effect on helping to reduce costs and make processes more efficient, and if so, they often follow suit.  

GIEDRIUS: It is obvious that AI creates or can create a lot of benefits for retailers, but what are the disadvantages, where can you slip up? Do you have any real examples? 

RIMANTAS: Just as with any other automation solution, AI works well when it is nurtured and maintained. If it is left to its own means there definitely could be some risks. But I don’t know of any actual negative examples yet.  

GIEDRIUS: Yes, and we face the same situation: ignorance of the fact that it’s not enough to just install a smart chatbot and that it requires constant work. We have many chatbots in place with AI for Lithuanian institutions that serve Lithuania as a whole, and we always try to communicate this: once a bot is launched, it must be maintained; not only technologically, as we do for our customers, but also in terms of content. Customers ask different questions, so you have to make sure that the chatbot responds properly and that the context hasn’t changed. For example, since the pandemic and the war, many laws, benefits, and so forth have changed, so the organisation that uses a chatbot has to adapt and provide it with fresh information. We have examples of success, such as the VMI chatbot being taught new topics and new subjects, and it services twice as many customers every year.  

You operate in 25 countries, so you may be able to objectively assess what Lithuania looks like in an international context with regard to the use of technology. When we work with our little language, we always bear in mind that all language technologies are first developed for the big languages, and our turn comes last. What about your technology solutions, they don’t depend on language and are probably developed in parallel everywhere, right? 

RIMANTAS: When comparing Lithuania with the Western markets, technologically, we are on par in the retail sector. Our consumers are spoilt; they want very high service quality, and retail businesses always want to meet these expectations. I have had a lot of visits to the United States, and I always like to watch the operation of self-service checkouts, for example. If it doesn’t work there, they simply stick on a note that it doesn’t work, while in Lithuania, they try to fix it as soon as possible, as otherwise customers will be dissatisfied. So our technology is really highly developed, even in comparison with, for example, the Scandinavian countries, which are the main markets of our group of companies. We’re the first to try out the latest technology here locally because our clients want to offer their customers the best shopping experience. 

GIEDRIUS: Well, to round up the conversation, let’s draw an image for the retail customer: so, if stores employ AI to the fullest at all levels, how does it change a routine grocery shop run, and what does it look like; how would it differ from the current situation? 

RIMANTAS:  Well, I’m sure, first of all, that physical stores are not going anywhere, as was once thought. E-commerce will take some share, but people will continue to shop in physical stores. For me, the store of the future is all about self-contained shopping when you arrive at your convenience, pick up what you need, and leave without any obstacles. Besides that, I think there will be technologies that we don’t have yet, which, for example, would allow us to try on clothes using an AI mirror without putting them on at all. I would also like shops to be more connected to homes, where the homes themselves see when a product expires and automatically order so that you don’t have to worry about it. Well, briefly, AI should ensure that your shopping run is smooth and effortless.