IoT and AI - Smart Management of the Whole Water Cycle

Can you explain what AI-based connected solutions SUEZ currently offers?

David – "At SUEZ, we offer a comprehensive range of connected solutions to support operators in the smart management of the water cycle. This offering covers the entire chain: from smart metering to data analysis, communication protocols and operating software. It is used both by our own customers – since SUEZ is also a water operator – and outside the Group, by public utilities, local authorities and international operators who work with our solutions.

The Internet of Things (IoT) allows us to capture vital information about networks in real time: volumes produced, volumes consumed, flow rate, pressure, water quality, etc. But this is only the first step. Artificial intelligence modules then use this data to determine the actions to be taken in the field – controlling valves, starting or stopping pumps, etc. – in order to optimise the cycle's operation."

What are the main benefits of these solutions?

"There are many. First, IoT enables us to detect abnormal situations more quickly and accurately: leaks, malfunctions, warning signs of failure. By categorising incidents, we can prioritise interventions and optimise resources. AI greatly amplifies this dynamic and allows us to go much further. Thanks to its computing power, pattern analysis and modelling, we are able to detect anomalies much earlier, anticipate breakdowns and refine maintenance operations. We are moving from corrective management to predictive management. Whereas older models were based on a few empirical variables, AI takes into account a multitude of criteria to produce more reliable and accurate diagnoses, truly harnessing the full power of data. It also allows us to simulate complex scenarios in order to plan preventive actions. Ultimately, it is a real performance lever, but also a means of drastically reducing water losses, securing networks, improving water quality and optimising resources."

Can you give us a specific example of how it is used?

"A prime example is the combination of our ON’connect™ smart metering solutions and our network modelling software, AQUADVANCED^®. By cross-referencing data from smart meters with data from water production plants and field sensors, we are able to use AI to pinpoint leaks before and after the meter, categorise them, and organise repairs in a much more targeted and efficient manner. In 2024, this system saved the equivalent of 8,900 Olympic swimming pools – or 33 million m³ of water – by reducing losses throughout the distribution chain. At the same time, we were able to identify areas of weakness and initiate preventive maintenance operations, rather than waiting for breakdowns to occur. But it is important to remember that AI is not everything: it is one component in a coherent whole that ranges from data collection to analysis and action in the field. Without proper data orchestration and trained teams, it does not produce results."

What are the obstacles that are still slowing down the adoption of these solutions?

"The main obstacle is the varying levels of digital maturity among operators. Some are very advanced and challenge the performance of the tools. Others are still in the early stages of digitalisation, with few processes and sometimes no precise mapping of their infrastructure. For them, a highly sophisticated solution is of no immediate interest: they need gradual support, with simpler offerings tailored to their context. That is why our range is modular and multitechnology: for example, we use the 169 MHz Wize protocol for high-performance remote readings, and other standards such as LoRaWAN or NB-IoT for simpler use cases (monthly billing, for example). And beyond technology, there is a real need for skills development. Digital technology is shifting the focus from jobs that are mainly focused on billing to jobs that are more focused on operational optimisation – water networks, flows, repair coordination. These changes must be supported by gradually training teams, because if you introduce a digital tool to a team that is not ready, it will not work. It is important not to try to move too quickly."

What are your plans for the future, the projects you are going to or would like to develop?

"Our priorities are focused on five areas:

• The widespread use of AI at all stages: design, operation and analysis. 
• The development of edge computing, which involves embedding intelligence as close as possible to the sensors for local data processing. This avoids the need to send massive amounts of (often useless) data upstream, limits the carbon footprint, and enables faster responses in the field. 
• Agentic AI, which is more automated and contextual, is currently being developed by our R&D teams to meet the specific needs of the water industry.
• Support for local authorities through consulting assignments, to help regions gradually integrate these tools, according to their level of maturity and in line with their actual needs.
• The prospect of a connected territory, where data on water, energy, transport and waste management interact with each other to automatically trigger the right actions. This may still seem a long way off, but the first experiments are already underway."

Conclusion from monreseaudeau: Artificial intelligence at the service of operational performance

Faced with the challenges of infrastructure renewal, resource constraints, and increasingly frequent and severe droughts and floods, water network management is undergoing a profound transformation as a result of digitalisation. The combination of field data, predictive models and decision-making tools now makes it possible to manage infrastructure with unprecedented precision. From leak detection and breakdown prediction to intervention prioritisation, AI brings a new level of responsiveness to increasingly demanding networks and strained resources. 

But this revolution cannot happen without a methodical approach. Computing power alone is not enough: it must be part of a coherent ecosystem, from sensors to field teams. This requires appropriate digital infrastructure, a gradual increase in operator skills, and technical solutions that are flexible enough to adapt to regions with varying degrees of maturity. 

Faced with these challenges, industry players are investing in modular solutions that can evolve in line with usage patterns. From simple automated meter reading to the deployment of increasingly sophisticated sensors and data processing, a wide range of tools is emerging to meet the needs of local authorities. 

The ambition is clear: to reduce losses, optimise resources and improve resilience, while preparing networks to enter a more systemic approach, where water interacts with energy, mobility and waste. Towards a truly connected territory.