"We also want get the skeptics on board and say, hey, we have an AI solution that's easy to use and still maintains your privacy!"

Simple, digital, quick and convenient: The increasing relevance of express delivery services, which bring their food packages directly to the door in ten to 15 minutes and are mushrooming in major German cities, also shows what customers are often concerned with these days. And while some people might now shy away from going to the nearest supermarket altogether, there are innovative ideas and ways to make the traditional supermarket visit more efficient. The new Berlin startup Nomitri, with the founding team of Trinh Le-Fiedler, Max Fiedler and Moritz August, has developed an autonomous and contactless self-check-out system based on Visual AI. This works without expensive camera installations, the setup of a cloud infrastructure and Internet connection, and incidentally avoids annoying queues. #ki_berlin spoke with CEO Trinh Le-Fiedler about their AI solution, the impact of the Covid pandemic and the future viability of German supermarket chains.

Hello Ms Le-Fiedler, how did Nomitri come about and what are you and your team working on?

We founded Nomitri in 2019. Moritz and Max (my two co-founders) and I had the idea for some time that we wanted to team up and build a deep-tech AI company in the embedded sector, yet we did not wish to start before we had a concrete product idea. It was important to all three of us that we develop a product that solves a real problem. In terms of experience, we complement each other very well. The deciding factor was then when Max was at a computer vision conference in the USA: There, he saw one of the first AmazonGo stores. That was our inspiration. Max came back to Germany and I immediately saw the potential of an AI solution for self-checkout. As a former director at a major e-commerce player, I know that brick-and-mortar retail can only benefit from digital solutions and fast data.

We see ourselves as an innovator in embedded deep-learning visual AI, which revolves around intelligent image recognition using neural networks. We develop state-of-the-art technology – research, development and the commercial side are very close to each other. There are not many companies in the world that do what we do: We are the first to have managed to compress deep learning algorithms in order to work on smartphones and other mobile devices; that is, they are "embedded" and "on the edge". Our technology is protected by a patent in process. The solution we are working on at the current time is our Nomitri app, a smart shopping assistant: A self-checkout solution for the smartphone that offers advanced features such as theft protection, personalisation and couponing – all with data protection in mind. Our solution is currently being piloted in Germany at Metro and Edeka Lüning.

What are the major benefits of your solution and how do you stand out from your competitors?

Nomitri has a unique position in the self-checkout solutions market. Solutions that are AI-based but rely on camera surveillance and sensors are very expensive – they can cost several million euros per store. In contrast, our solution only requires a mount on the shopping cart, and retailers pay a fee to use the app – a clear cost advantage. Customers simply use their smartphones. Traditional scan & go solutions are similarly priced – but they don't offer AI benefits such as theft protection and personalisation. As one can see: Nomitri certainly has a unique position, as we combine the "best of both worlds” so to speak. The most cost-effective solution and clear benefits from AI.

On the topic of the Covid pandemic, contact reduction and self-checkouts in the store: Did this also play a significant role in why your solution has found such resonance?

We created Nomitri in 2019, before Covid became an issue. The pandemic is most certainly driving innovation, and not just in our sector. Retailers are looking for new solutions to bring people back into stores. In-store innovation can create a unique shopping experience, and we believe it can help retailers appeal to people who currently prefer to shop online. In general, online shopping offers many conveniences that make shopping in a store seem costly and cumbersome. What the pandemic has shown is that customers want to shop quickly and without contact above all else – and that's exactly what our smart shopping assistant makes possible. Moreover, after Covid, the digital transformation will play an increasingly important role in brick-and-mortar stores.

To what extent did data protection considerations factor into the development of the AI shopping assistant?

We do things differently than well-known players in the market. We rely on non-invasive technology. When customers shop with our app, they mount their smartphone on their shopping cart. No faces are recorded. In addition, video data never leaves the customer's smartphone. Well-known solutions, such as AmazonGo, rely on camera surveillance and sensors – we want to get away from that, or didn't want to go there in the first place. We have also been able to learn from each other's mistakes and do better from the start. Artificial intelligence is still subject to many prejudices. Socially, we are still at the beginning in terms of AI acceptance, but I believe that will change when AI actually accompanies us in everyday life in various areas. We also want get the skeptics on board and say: Hey, we have an AI solution that offers many benefits, that's easy to use and still maintains your privacy!

Is the retail sector in Germany and the rest of the DACH region open-minded enough to embrace the new technology and its possibilities? What has been your experience?

In fact, it makes a big difference: In our experience, the retail industry in Germany is conservative. Many retailers show interest, but a bureaucratic gauntlet then often follows until a demo even materialises. The digital transformation has been neglected in Germany for too long, and the retail sector is no exception. People are afraid of alienating older customers. Retailers in other countries are more open-minded and are approaching us on their own: They have experienced that, on the contrary, digital solutions can be used to address new target groups and increase sales. But that's where our solution comes in: With Nomitri it is possible to "transform" a store step-by-step. There will still be normal checkouts, but also the option to pay with the app. What's more, the app works on almost any smartphone – most customers already own one, so there's no need to get used to a new system or the like.

According to the analysis of one of the largest European food retail chains, shoppers using the competitor product, MishiPay Scan & Go, spend 14 percent more in supermarkets. Does this confirm your entrepreneurial thinking?

According to our analyses, it goes even further: A study by McKinsey has shown that purchase recommendations already account for 35 percent of sales at Amazon. Thanks to the use of MishiPay, the increase in sales can probably be attributed to the conveniences that come with its use, such as time savings and "fun shopping". Whether these results will be repeated remains to be seen. We expect to see even larger "basket sizes" (that is, how much a person scans during a purchase) through the use of our app, as we also offer personalised recommendations and couponing. Furthermore, a study by 451 Research and Adyen shows that American retailers lose more than $37 billion annually from customers not even entering the store or abandoning their purchase due to waiting times. This is also where we come in and offer a solution to this problem – thanks to self-checkout.

Beyond that, what else are you planning with your partners for further ventures in the field of AI? Can the AI concept be transferred to other sectors? If so, which ones?

Our partners are market leaders in their segment. Medion and Gebit support us primarily with their expertise in the retail sector and thus also when entering the market with our intelligent shopping assistant. Variscite and Phytec are leading System-on-module producers who support us with the necessary hardware. Collaborations in other sectors are also conceivable here, for example in the construction of intelligent drones or other industrial applications in the future. Visual AI that works in embedded form and "on the edge" has a great deal of potential applications: In principle, one can make any camera "intelligent" and thus its use much more efficiently. Weather monitoring, surveying and metering would be possible here, to name just a few possible applications.

We are also planning a cooperation with a global retail outfitter, but this may not be announced yet. We're also planning to bring AI to brick-and-mortar stores via a different channel.

What does the Berlin AI ecosystem offer you as a company? What makes the location so attractive for you?

Berlin – as the capital city – serves as a great attraction for highly qualified employees, including those from abroad. This is also evident in our team structure: People of ten nationalities work at Nomitri – it's remarkable, after all, we number only 14! Berlin continues to grow as a startup centre.  I believe that it would be more difficult to find the right personnel for us at another location. Thanks in part to our exciting mission, we have managed to recruit an excellent team, some of whom have many years of international work experience in the AI field.

Of course, the AI environment in Berlin also offers some advantages: In addition to networking and exchange with companies, there are also opportunities for cooperation. In addition, many investors are also based in Berlin.