Interview
Optimizing Urban Cleanliness: Andréas von Kaenel on Pioneering AI-Driven Solutions with Cortexia Ahead of IFAT 2024
Published on May 7, 2024
Andréas von Kaenel CEO and Co-founder at Cortexia
Street cleanliness is a significant issue for city leaders. It can really affect the way a city is perceived. A dirty city may scare away tourism and business and can even cost politicians an election. It's also very important for public health, as dirty streets can block drains, leading to flooding, or attract rats, which are known to spread diseases. Cleanliness operations also require vehicles and manpower and can represent a significant share of cities' budgets. Considering the challenges, Kurrant Ventures invested in September 2023 in Cotexia, an AI-enabled smart cleanliness company that has made optimizing cities' cleanliness and operation costs its mission.
Ahead of IFAT 2024, Munich's trade fair for water, sewage, waste and raw
materials management,
we spoke with Andréas von Kaenel, CEO of Cortexia, to discuss the importance of cleanliness
efficiency in cities.
During our interview, Andréas, an engineer who was part of the Ariane 5 rocket’s first test
flight, saw the birth of the solar industry from the inside and a believer in a circular and
sustainable economy, told Kurrant about the importance of street cleanliness with Cortexia as an
example, a company that is leading the way in the Clean Space Race.
Kurrant: Andréas, IFAT is right around the corner, and waste management is a major topic of the fair. So the first question has to be, what are the main challenges a city faces when it comes to cleanliness and how can digital solutions help municipalities face them?
Andréas: I think the first challenge is that it's very subjective. Also, for the cleaning organization, they’ve kind of had the same way of working for many, many years. And bringing these new technologies gives them tools to improve their job. We created the company because I met someone who was consulting for cities that wanted to know how many street sweepers they needed to keep cities clean. We were always talking about means but never results. As soon as there was a new machine (street sweeper) in the city, they were using it 100%. But they never checked what the quality of the result was. And so, we found out that to answer the question ‘How many machines do I need?’ you have to answer the question ‘What is the level of quality I want to achieve?’. And then to be able to manage or to optimize it, you have to measure it.
Kurrant: So the idea of Cortexia came from a need you and your partner at the time saw to measure the quality of the cleaning work. You then designed a solution. Your company mounts cameras and a box equipped with artificial intelligence on street sweepers to identify and classify street rubbish, to understand how clean or dirty streets are, and therefore, manage street sweepers more efficiently. Tell me, what actions can cities take thanks to gathering this data? What are the benefits of investing in smart cleanliness?
Andréas: Because it's a lot of data, we try to simplify the data for the customers. First, we introduce this Clean City Index. So it's kind of a grade you can give. So it's between 0 and 5. You gather all this data which says, 'Okay, you had so many cigarettes, so many bottles...'. What we do, we translate it into colours (on a map). The city defines a level of cleanliness they want to achieve. So all the streets which are within this quality level, appear in green. The streets which are too clean appear in blue, meaning they are using (too many) resources, costs and the environmental impact… So those sweepers will be better used in the red streets, which are the dirty ones. So you can check on a certain time and see that, in some areas, some quarters, you go too often and sometimes you do not go often enough. This is the first way to optimize, adjusting the cleaning frequencies. The second one is you can even adjust the routes you are taking to clean. One of our customers, the city of Utrecht, mentioned that 70% of the time they were cleaning streets which were already clean. You can also try to avoid littering. We can superpose the map of the cleaning with the map of the trash bins, then, you see directly there are regions where you have a lot of trash and no bins.
Kurrant: And where are cities investing in this type of solution seeing the ROI?
Andréas: The City of Basel has a very ambitious climate plan and they want it to be carbon neutral by 2025. That means, for the cleanliness department, they have to change all their vehicles to electrical sweepers, they are twice as expensive as the other ones. So it was a real problem for the manager. And he said to his team 'We really have to reduce the number of vehicles because we cannot afford it'. They have compared the cleanliness maps with the use of the machines. They have reduced the number of machines by 15%. We had almost a similar experience in Utrecht and also in Geneva. In Geneva, they could clean at a higher level of cleanness with two machines instead of three, saving 20% of machine hours. A sweeping machine uses five times more diesel than the other trucks.
Kurrant: Your solution is possible thanks to AI, as Vision AI allows you to classify all kinds of litter, even small items like cigarettes. How important is the role of AI becoming in the industry?
Andréas: In a way it's disruptive because of what we've already seen today just with the detection. To be able to measure a cigarette, and count the cigarettes at the speed of 30km/h with a car driving, it's disruptive, what we see now. We have made a next step. The question is not only to measure but to optimize. And today we can optimize manually. So we have learned how to optimize based on data, but, in the future, we will do it automatically. So when we measure the level of cleaning as we can, plan the cleaning and adapt the cleaning from the days before etcetera based on historical data, maybe also on prediction. We have taken the step with academic partners and also with the City of Basel. Today, the results were quite good, but not good enough for industrial use. I think the next really disruptive thing is when we have a tool to manage the resources. We were the first startup where EPFL (university) developed deep learning algorithms. So of course they already mastered deep learning for their research. But it was the first industrial use case. And from 2016 till we had a real reliable industrial daily use solution, it took some time. So of course, disruptive is if you want to have something which is in daily use, it takes some time.
Kurrant: You recently raised 2.4 million euros in funds from Remondis Digital Services, Kurrant Ventures, Capital Risque Fribourg as well as Swiss VCs Bloomhaus and Spicehaus Partners. What was your experience like? What's your advice to a start-up looking to raise funds?
Andréas: When we started raising, we started the fundraising too early. We were not really ready. Fundraising takes really a lot of resources, time, energy. So you really have to be ready and keep the period as short as possible, because it will divert you from your main business. Of course, for smart city (companies) it is very difficult. In our case, in the second round, we had strategic investors. You normally do it the other way around. So you start with VCs and then you have strategic investors. But we had VCs coming in the third round because VCs don't like startups in the smart city area or the tech area if your customers are public customers. So the life cycle or the acquisition cycle of public customers does not fit the criteria of VC funds usually.
Cortexia’s successful fundraising round proves how street cleanliness is becoming more and more a focus point for cities. Smart cleanliness allows cities to improve services while reducing costs, so solutions like that of Cortexia, which helps cities become more efficient in their cleaning, are set to become a trend in environmental services, one of many we’ll see at IFAT 2024, where Cortexia will be exhibiting next week
Interested in our investment strategy?
Visit Kurrant Ventures© Kurrant. All Rights Reserved.