Distinctive Group

Kevin Sheeran, Technical Specialist, Suez Aqua Enviro, investigates how wastewater utilities can make best use of AI.

Artificial intelligence (AI) will become ubiquitous in our daily lives, powering services from digital assistants to personalised content recommendations. Many industries will benefit immensely from harnessing AI, but it is imperative that we begin to consider what that will look like within the water industry. AI could enable smarter management of our vital water infrastructure. As utilities develop business plans for PR24 and AMP8, AI adoption should be a strategic priority to enhance operational efficiency, meet performance commitments cost-effectively, and unlock new value.

However, AI represents a paradigm shift, enabling systems to learn, reason, adapt, and function with higher autonomy.

AI vs. current analytics approaches

Many wastewater utilities already utilise some data analytics and machine learning. However, AI represents a paradigm shift, enabling systems to learn, reason, adapt, and function with higher autonomy. Instead of just pattern recognition within data, AI incorporates contextual understanding and judgment – delivering more impactful insights and recommendations. An AI virtual assistant, for example, can have an intelligent dialogue with operators. AI’s continuous learning and optimisation capabilities are ideal for managing highly variable and complex processes like wastewater treatment. AI is already being trialled as a tool to predict leaks within the sewer networks. Overall, AI unlocks far greater value than current analytics approaches.

Treatment process optimisation

AI can optimise treatment plant performance through data-driven process adjustments, predictive control, and anomaly detection. Sensors can provide rich real-time data on parameters like flow, temperature, pH, dissolved oxygen, etc, while AI could analyse this data to model process behaviour, identify optimisation opportunities, and enable automated adaptive control. This would reduce energy and chemical use, ensure permit compliance, and lower carbon footprints – aligning with Ofwat’s focus areas. AI techniques, like computer vision applied to microscopic imaging, could automatically assess sludge conditions and microbiological health. Chatbots could serve as smart virtual assistants for operators. AI-based treatment optimisation could greatly improve efficiency, resulting in cost and time savings.

Predictive maintenance

Equipment failures account for significant wastewater operational costs. AI predictive maintenance would analyse historical sensor data to accurately forecast asset deterioration and proactively schedule interventions, consequently minimising downtime and avoiding catastrophic failures through early diagnosis and root cause analysis. AI also has the potential to enable condition monitoring through automated techniques like vibration analysis.

Optimising anaerobic digestion

Anaerobic digestion (AD) of sewage sludge is critical for waste reduction and biogas energy recovery. However, AD is a complex microbial process affected by many factors, which could be optimised by the use of AI techniques. By analysing digester sensor data on temperature, pH, gas production, etc., AI models could determine ideal operating conditions and predict biogas output. This would enable tighter real-time control and dynamic adjustments to maximise gas yield. Additionally, AI would be able to detect potential process instability early, allowing pre-emptive intervention. Diagnosing AD process upsets quickly is crucial to avoid extended recovery. Overall, the insights that AI optimisation has the potential to provide could significantly improve AD efficiency, gas production, and resilience.

Wastewater epidemiology

Wastewater-based epidemiology (WBE) powered by AI would unlock major public health value from sewage, through detection of pathogens, chemicals, and health biomarkers in wastewater. Thus enabling real-time monitoring of disease outbreaks, new substance flows, and community health trends, AI-powered WBE would be a vast improvement upon current models. WBE is already becoming vital for public health resilience and emergency response, but with AI bio-surveillance, wastewater and public health authorities would be able to more efficiently respond to shifts in data.

Integrating AI and Metagenomics

Metagenomic analysis of wastewater microbiomes is still in its infancy, but generates vast amounts of data on microbial populations. AI is potentially necessary to extract useful insights from these complex datasets. High-throughput metagenomic sequencing combined with AI would enable rapid functional profiling of microbial communities. AI could track community dynamics to optimise biological treatment and sludge management. Overall, integrating AI and metagenomics provides unique value for characterising and enhancing biological wastewater processes.

Driving Net Zero progress

Achieving net zero operations is a major goal for wastewater utilities: the use of AI could accelerate sustainability initiatives and decarbonisation efforts. By optimising treatment processes and equipment efficiency, AI would reduce energy usage and lower carbon footprints. AI-enabled predictive maintenance could also minimise greenhouse gas emissions from equipment failures and unplanned downtime. Additionally, enhancing anaerobic digestion with AI increases biogas production, providing renewable energy to offset grid consumption. AI data analytics would support utilities in tracking, reporting and reducing emissions. Overall, AI has the potential to deliver significant environmental benefits, enabling wastewater companies to make major strides towards net zero.

AI should be positioned as empowering human operators, not replacing them.


As AI adoption grows, all industries must prioritise ethical implementation. AI systems should be transparent, explainable, and unbiased. Operators should be able to understand and validate AI recommendations. Workforce concerns about AI threatening jobs must be proactively addressed through change management and skills training. AI should be positioned as empowering human operators, not replacing them. AI should ultimately benefit communities, by improving public health, environmental quality and service affordability.

The path forward

To fully capture AI’s potential, utilities will need to invest in sensors, data platforms, automation, and workforce skills. Progressing AI adoption in an ethical, unbiased, and transparent manner will be critical. The innovation funding being made available in the coming years offers an opportunity to accelerate AI deployment. Through strategic AI implementation, utilities can achieve new heights of operational excellence and sustainability, while also providing immense societal value.

A noteworthy initiative in this realm is the DARROW project, initiated in 2022. This project focuses on training AI models to analyse sensor data from wastewater treatment plants. These AI models provide invaluable recommendations, aiming to reduce energy and chemical usage by 20%, minimise sludge production by 50%, and decrease emissions by 20%. Additionally, the project emphasises enhancing energy and nutrient recovery processes, aligning with the broader goal of sustainability.

Looking ahead, the future appears promising, as AI continues to evolve and positively influence the world. Through optimised management facilitated by AI, we can envision a future where water resources are utilised efficiently, benefiting both the environment and society at large. The strides made in AI-driven wastewater management not only signify progress in technology but also underscore our commitment to safeguarding our planet’s most essential resource.