CEO of Kheoos
Kheoos helps to extend the life of maintenance parts, while meeting today’s environmental and economic challenges.
… our ambition to promote a more sustainable industry and our innovative use of AI enable us to attract talent passionate about technology and the ecological transition. 
Could you tell us about the origins of your company, describe its activities and the special features of the technology developed by Kheoos?
Kheoos was founded by three people: Emmanuelle Paris, Guillaume Lepelletier and myself. All three of us had the opportunity to work in large companies, where we observed the difficulties associated with the accumulation of unused industrial maintenance parts and the lack of effective solutions for managing inventories. Dormant industrial inventories are estimated at 30 billion euros in Europe. This represents a major challenge for industry in terms of cost, sustainability and inventory optimization.
We decided to create Kheoos in 2018 to offer an innovative solution in this area. Our mission is to make industry more sustainable by facilitating the reuse of maintenance parts. We have developed a technology platform, mykheoos, that enables companies to manage and reuse their stocks of dormant parts rather than throwing them away.
What sets Kheoos’ technology apart is its ability to analyze and add value to parts dynamically. We integrate data science, artificial intelligence and big data tools to create an intelligent ecosystem where parts are identified, sorted and redistributed through an open platform that connects companies, reducing waste and the costs associated with purchasing new parts. In this way, Kheoos helps to extend the life of maintenance parts, while meeting today’s environmental and economic challenges.
How is artificial intelligence integrated into Kheoos’ processes? What is its added value?
Artificial intelligence is at the heart of innovation at Kheoos. It is integrated at several levels of our processes to maximize the efficiency and accuracy of our industrial parts management.
AI is used first and foremost to create an advanced industrial reference recognition tool. This solution automatically identifies parts in our customers’ inventories from a database of 85 million brands and references, verified and labeled. This technology cuts the time needed to identify parts in an inventory by a factor of 10, eliminating up to 30% of duplicates in databases on average. This makes inventory management and part reuse much faster and more accurate.
We have also integrated language models (LLM) as an overlay in some of our business tools. This type of AI program can automatically generate missing data or perform translations. This optimizes our internal processes and improves the fluidity of data exchanges.
What’s more, integrating LLM bots into our development tools has enabled us to automate certain technical tasks, reducing development times and boosting our team’s productivity.
Finally, thanks to the use of tools such as ChatGPT and Copilot, we have automated office tasks such as document drafting, information retrieval and internal communications management, enabling our staff to concentrate on higher value-added missions.
What are the skills and expertise of Kheoos employees?
Kheoos currently has 20 employees. We bring together a wide range of complementary profiles, combining a vast array of IT skills from web developers to cloud architects, as well as operational research experts to develop optimization solutions. All this is complemented by skills specific to the development of artificial intelligence (data engineer, data scientist and data analyst). We also have business experts with extensive experience in industrial maintenance, with the aim of maintaining a strong link with our customers’ businesses.
Traditional IT skills ensure the cohesion and fluidity of our platform, and guarantee the scalability and security of our systems, particularly when it comes to managing the large quantities of data we process.
In AI-specific skills, data scientist profiles play a key role in the design and optimization of our predictive models and artificial intelligence algorithms, enabling us to maximize the efficiency of our solutions. Data analysts process and interpret data to extract useful and relevant information, contributing to strategic decision-making. We also need operations research specialists to develop efficient search algorithms that optimize inventory management and duplicate detection.
Like many technology start-ups, we face recruitment challenges, particularly for profiles specializing in data science and operations research. These skills are in high demand, and the market is quite competitive. However, our ambition to promote a more sustainable industry and our innovative use of AI enable us to attract talent with a passion for technology and the ecological transition.
What are the conditions for success in developing your project? What are the risks of using AI?
The success of an artificial intelligence project depends on several key factors. First of all, good data quality is essential. The more reliable the data, the more effective the AI models. Secondly, it’s crucial to ensure smooth integration of AI into our existing processes, while guaranteeing team involvement in the deployment of AI tools.
AI can be an extremely powerful tool, but it carries risks that are important to consider.
First of all, AI models generally achieve performance rates of 90-95%. However, an error rate of 5% is still too high, especially when it comes to important decisions. On the other hand, LLM models, while impressive, can suffer from so-called “hallucination” phenomena, where they generate erroneous or inconsistent results. This limits their reliability and unsupervised use.
At Kheoos, to overcome these limitations, we have implemented targeted human controls throughout the process. This enables us to guarantee an accuracy rate of 99.5% in our AI tools. These human checks are crucial to ensuring that AI delivers value while minimizing the risk of error.
How would you sum up the last two years? How do you see KHEOOS evolving in the coming years?
The last two years have been marked by massive investment in our technology. Last October, we reached a key milestone with the launch of mykheoos, our SaaS platform. This solution enables our customers to subscribe to and instantly access a full range of services: consolidating parts databases, putting their dormant inventories up for sale, proposing purchases from the community’s dormant stocks, and anticipating parts obsolescence by tracking manufacturers’ declarations.
In 2025, we will continue our R&D efforts with a major European project, in consortium with industry leaders such as Michelin, John Cockerill, and European academic institutions, focused on virtual inventory sharing. This project will enable us to further optimize maintenance parts management on a European scale.
Today, the Kheoos community includes 90 warehouses in 13 European countries, with major players such as Michelin, EDF, Sanofi, Poma, Valeo and many others.Over the next few years, our aim is to significantly expand this community and become the European reference in the B2B circular economy for industry, particularly for maintenance parts.Our ambition is to transform the way companies manage their dormant stocks, and to reinforce the positive impact of the circular economy in the industrial sector.
