Water operators and decision-makers in the water sector today have access to more detailed and (near) real-time data than ever before thanks to modern digital tools — like smart sensors, SCADA systems, GIS platforms, and digital twins. Whether it's monitoring flow rates and water quality indicators at a wastewater treatment plant, or detecting pressure drops across a city’s water distribution network, digital systems are transforming daily operations.
However, collecting raw data is not enough. What truly matters is how that data is interpreted, understood, and used. The complexity of transforming raw data into actionable insight—especially in the wastewater sector—is well explained here.
Enhancing the value of collected data is where advanced data analysis tools and modeling approaches truly make a difference. From well-established mechanistic models to AI-powered data-driven models, and increasingly popular hybrid models that combine both, these tools support better decision-making at every level.
AI systems are not meant to replace operators in the control room — they empower them
Prediction tools — such as model predictive control (MPC) — also allow operators to anticipate problems and act before they arise. With the added power of artificial intelligence (AI) providing real-time insights, the water sector is stepping into a new era of smarter, more proactive, and more resilient management. At the same time, new challenges such as energy conservation and resource reuse, aligned with circular economy principles, call for strong collaboration between people, technology, and research.
In this recent opinion paper, the authors highlight the significant potential of machine learning (ML) for wastewater applications — considering hybrid modelling in particular very efficient — but also caution against applying ML indiscriminately to all data-related challenges. They emphasize the importance of keeping ML models interpretable and trustworthy by avoiding unnecessary complexity. Crucially, the paper underscores the human factor: training operators in new data analysis tools is essential to support and enhance their role in daily operations.
While technology is evolving rapidly, people remain at the heart of water operations. AI systems are not meant to replace operators in the control room — they empower them. Data-driven insights don’t eliminate the need for human expertise; instead, they provide operators with better tools to understand their systems, respond quickly, and make smarter decisions.
True progress happens when innovation and research come together — when researchers develop advanced models and methods, and innovators turn those ideas into practical tools that withstand dynamic, real-world operating conditions. The real strength of the digital water transition lies in the collaboration between skilled operators, smart AI systems, and continuous research innovation. This synergy between people, technology, and science is what’s building a more resilient and efficient water sector for the future.
Smarter wastewater treatment plants (WWTPs)
In modern WWTPs, AI and modeling tools are becoming powerful allies for operators. Advanced AI-driven systems can monitor incoming pollutant loads, detect anomalies, and classify the severity of events to scale their potential impact. They can also predict biological process behavior, recommending real-time adjustments to aeration or chemical dosing.
Instead of replacing skilled operators, these tools act like an additional set of eyes, helping the controller anticipate issues and make quicker, more confident decisions. Claire Mathieu, Digital Solutions Director at SUEZ, captures this spirit: "We transform millions of data into concrete action," highlighting how AI supports operators in making informed decisions.
Hybrid models — combining real-world process knowledge with data-driven insights — are increasingly used to stabilize effluent quality, minimize energy use, and detect anomalies in monitoring data and related operational problems early-on. See this publication of a recent development of a hybrid model for a biofiltration system in a large-scale WWTP in France as part of the innEAUvation applied research program (SIAAP).
Another example is the Real-time AI-Driven Decision Support System (RADS) developped by DC Water. RADS is a web-based platform that processes a high stream of real-time data and supports cost-efficient, safe, and secure decision-making processes at water treatment facilities. It supports multiple WWTP tasks; from cybersecurity and effluent quality prediction to optimizing treatment during extreme weather events. It helps operators manage large volumes of data, monitoring such data and interpreting underlying hidden information in a timely manner, which is crucial for decision-making and ensuring safe, secure, and efficient operation at a facility. This tool allows to improve effluent water quality, enhance energy efficiency, and build operational resilience. Rather than replacing operators, the RADS amplifies their decision-making power, especially when it matters most (during a-typical events).
The implementation of AI-driven decision support systems has already shown improved operational performance and environmental compliance. For instance, see the research work of the modelEAU research group at ULaval (Quebec, Canada), where Jeffrey Sparks focused on the development of advanced nitrogen removal control systems in a full-scale WWTP plant using hybrid models.
Optimizing desalination plants
Desalination is energy-intensive and sensitive to operational conditions like pressure, membrane fouling, and salinity changes. AI-driven predictive tools, including model predictive control (MPC), are helping operators fine-tune systems to maintain performance while reducing costs. By analyzing trends in intake water quality and equipment behavior, these tools can suggest optimal setpoints or maintenance actions before issues arise. For example, early warnings about membrane fouling allow teams to intervene proactively, reducing downtime and extending equipment lifespan. A compelling example is MEMBoard, a web-based software to document and report O&M actions on membranes in a desalination plant, developed by SUEZ's research centre CIRSEE. The tool enables operators to track membrane module performance and predict replacements and failures.
Training & trust: Equipping operators for the digital era
Water utilities should prioritize regular training of operational technology (OT) staff operating industrial control systems (ICS), such as the SCADA systems, and the infrastructure data associated with their networks. This training is vital for ensuring business continuity and securing operations. As advocated in an earlier blog on the importance of in-house operator training in the water sector, training programs should integrate AI, machine learning algorithms, and simulation models to provide operators with cutting-edge, interactive training experiences.
This way, digital twins turn “what‑if-scenarios” into hands‑on practice allowing operators to be able to practice in a safe environment all kind of calamities; rehearse storm surges, equipment failures, or process tweaks inside a virtual reality of their plant. Digital‑twin platforms can stream real‑time SCADA data into a dynamic model allowing accurate predictions of different KPIs. Specialized operator‑training simulators use the same underlying process models engineers trust for design. Trainees can try “what‑if” scenarios (for example high ammonia, blower trip, power outage) without risking operations and running outside of compliance limits. These simulators will allow to optimize decision‑making that can translate directly to lower energy and chemical consumption.
Operators are far more likely to act on AI recommendations when they understand why the system is proposing a certain change. Explainability of a proposed action is becoming an integral part of a model, where instructions for important actions are communicated to the human operator accompanied by short and clear explications (e.g.: “Reduce airflow because oxygen uptake rate is falling!”) instead of displaying only recommended set‑point changes. Data-driven models take on the task of interpreting what’s happening and suggesting the best recommendation(s), assisting operators making informed decisions.
As AI tools evolve, continuous upskilling keeps operators and decision makers in the loop with short, recurring “micro‑trainings” (eg: 15‑minute dashboard refreshers) together with monthly “model status” reviews, and annual simulator drills. This steady pace lets staff absorb new features gradually and gives them a say in refining alarms, KPIs, and report formats; turning technology adoption into a two‑way conversation rather than a top‑down mandate.
Together, immersive training environments and transparent AI logic aren’t replacing operators — they’re raising their game, turning years of practical know‑how into faster, data‑backed decisions trusted by management.
Conclusion & perspectives
Real digital progress happens where lab insights meet operator‑room reality. Practice‑driven and data-driven innovation will lead to digital tools shaped side‑by‑side with the operators who will use them; this will turn brilliant programming code into everyday value. A recent example: SUEZ and France’s CNRS have just signed a five‑year strategic partnership to fast‑track sustainable‑water breakthroughs from the lab to the field, underscoring how industrial know‑how and academic excellence accelerate impact together.
When water utilities provide access to their databanks and define their pain‑points, and academics bring modelling depth and fresh ideas, co‑development accelerates: prototypes iterate faster, trust builds sooner, and solutions launch “operator‑ready.”
Key take‑aways
- People first: AI augments, not replaces, operator know‑how.
- Explainability wins trust: transparent models are acted on and not ignored.
- Practice‑driven & data-driven R&D: pilot studies at the plant and the application of digital twins can close the gap between theory and the operators plant floor.
- Shared journey: utilities, software developers, system design engineers and research centres create the strongest results when they design, test, and refine together.
The smartest water systems aren’t just digital — they’re human‑powered, data‑fueled, and built in partnership.
