The Center for Data Innovation spoke with Marjorie Darcet, co-founder and CEO of Lixo, a Paris-based software company that uses computer vision and AI to optimize waste-hauling practices. Darcet spoke about how Lixo’s computer vision technology helps reduce the risk of hazardous materials in waste hauling and how the industry can use data to make waste management more efficient and environmentally friendly.
Martin Makaryan: What does Lixo offer?
Marjorie Darcet: Lixo provides smart waste collection solutions. We install small devices with cameras inside waste hauling trucks, oriented towards the hopper where the bin is emptied. These devices run computer vision algorithms that can detect and classify different waste items in real-time. We then send this data to our clients, which are primarily private waste management companies, like Veolia in France or municipalities that choose to carry out waste collection and management themselves. We are expanding into other European and North American markets. We provide data through application programming interfaces (APIs) or custom dashboards, providing them with detailed information about the waste they are collecting.
Makaryan: How does your technology improve waste management?
Darcet: Our tools bring value by generating, processing, and analyzing data that waste haulers generate. We can detect dangerous items like medical waste in real-time, preventing potential accidents or environmental hazards. We also monitor waste quality, ensuring that haulers properly sort and direct specific types of waste, like bio-waste, to the correct facilities. This data-driven approach allows them to prevent contamination, especially in composting processes where plastic contamination can lead to pollution. In the long term, our data helps clients target their communication campaigns to improve sorting practices across their service areas. We can also identify specific clients of waste management companies with whom we work, like restaurants or cafes, that consistently contaminate waste streams, allowing our clients to address these issues directly. Reducing sorting errors during waste collection saves money for both the private waste collection companies and the taxpayers who ultimately pay for these services.
Makaryan: How does Lixo use AI?
Darcet: AI is one of the pillars of our system. Our AI algorithms analyze images that the small devices installed on hauling trucks capture to identify and classify waste items. While AI is crucial, it’s important to note that the hardware component—our image capture devices—is equally essential. And while our AI team is relatively small, it is highly focused on developing effective solutions for our specific use case. For example, we continuously use the captured images and data, whenever possible, to identify new materials, like new types of plastic used in food packaging, and train our models to fine-tune them for more local contexts. This customization is especially important as we are expanding into other markets.
Makaryan: What has been your biggest challenge as an innovator?
Darcet: The biggest challenge for me has been the business aspect of our operations—convincing potential clients of the value that a data-driven approach will bring to waste collection and processing. Because of the nature of the waste hauling industry, we are constantly dealing with large companies and public administrations, which can sometimes be slow to act, and that can have a negative impact as well.
Makaryan: What impact do you hope to achieve with Lixo in the future?
Darcet: Our vision is to become the reference company for waste understanding. We believe that waste is a resource, but it’s largely unknown and poorly managed. By providing detailed, real-time data about waste composition and quality, we want to optimize the entire waste management value chain, from collection to recycling. Beyond effectiveness and cost savings, a data-driven approach has significant environmental benefits as well. For example, by improving the sorting of bio-waste and preventing contamination, we can reduce microplastic pollution in composted materials. We are also expanding into other types of waste beyond residential, like commercial and industrial waste, and potentially construction waste in the future. Ultimately, we want to create a comprehensive platform that can analyze and provide insights on any type of waste, helping to make waste management more efficient, cost-effective, and environmentally friendly.
