 Making buildings smarter is one of the big challenges of today's companies to improve their efficiency by saving time, money and the environment. In this context, knowing the number of people in a given area or waiting in front of a room or restaurant enables us to have an overview of the attendance in real time. This gives us smart building insights that enables it to estimate the waiting time, predict the amount of people and as a consequence better manage services and energy. The people flow counting sensor developed by Schneider Electric in partnership with ST Microelectronics enables the counting of the number of people and the detection of whether they are crossing a virtual line in both directions. Using a large field of view and a small resolution thermal sensor, this prototype can count in real time and with a high level of accuracy the restaurants attendance while running on a standard STM32 microcontroller. Thanks to the artificial intelligence algorithm embedded on the STM32 microcontroller and the use of a thermal infrared technology, this solution respects people privacy by design and is easily deployable in any building. But let's have Maxime and Miguel tell us more about this demonstration. Our people flow counting demonstrators runs a neural network on the STM32 H7 microcontroller. This powerful chip enables us to process all the thermal images anonymously in the sensor itself directly without the need of any kind of cloud connectivity and this thanks to ST Microelectronics. It has been a very exciting time working together Maxime. Thanks to the STM32 QBI tool that is a tool that converts pre-trained neural networks quickly and efficiently. We have been capable to iterate quickly and narrow down to the best and more optimized neural network architecture to fix this problem. We would like to believe that thanks to this gain in productivity your team has been capable to focus on the rest of the application. Absolutely. We even enjoyed a performance increase during the life of this project simply by upgrading STQBI to the latest revision. Your team made a huge effort making neural network compatible to the world of MCU. This allows us to develop rapidly a sensor showcasing the future of digital attendance monitoring. In the STI team we are always excited to see how AI technologies can extend the capabilities of the standard and general proposed microcontrollers. This will really allow new type of products to hit the market thanks to the fact that they will come at the right price point. Indeed. We can easily imagine how this technology will accelerate digitization for smart building, for schools, hospitals, retails, offices and any smart building of any kind in this world.