Water Utilities See Clear AI Use Cases, but Data Foundations Remain the Barrier

Artificial intelligence is moving rapidly up the agenda for water utilities, but the sector’s readiness depends less on the algorithms themselves and more on the quality of the systems beneath them.

That was the central message from recent discussions at Ozwater 2026, where operators, engineers, analysts and asset managers explored how AI could support core utility functions. According to an article by Tim Westphal, ACT Branch Committee Member at the Australian Water Association and Account Executive for Public Sector and Utilities at Atturra, the sector has already identified where AI could add value. The harder question is whether utilities have the digital foundations to support it.

Water utilities are not short of data. Many have been collecting operational, customer and asset information for decades. The challenge is that this information is often fragmented across SCADA platforms, asset management systems, customer systems, spreadsheets, partner reports and legacy databases. In that environment, AI can expose operational weaknesses rather than solve them.

AI Interest is High, but Readiness is Uneven

The most attractive AI use cases for water utilities are practical and operational. They include earlier leak detection, more accurate demand forecasting, better asset performance monitoring, optimized treatment processes, and improved field crew coordination. These applications could help utilities reduce water losses, lower energy use, improve service reliability and respond more effectively to extreme weather events.

However, the source article notes a consistent gap between ambition and readiness. Some utilities still rely on manual reporting processes, with performance data assembled across multiple delivery partners and dozens of separate reports. In one example cited, reporting operated on an eight-day lag. Elsewhere, early AI pilots had stalled because the data feeding them was unclassified, poorly governed or scattered across systems that did not communicate effectively.

This matters because water operations often involve decisions with immediate public health, environmental and service implications. If data is incomplete, delayed or difficult to verify, utilities may be reluctant to rely on AI outputs for operational decisions. In practice, that can limit AI to narrow pilots rather than enterprise-wide tools.

Connected Data is Becoming a Sustainability Issue

For net-zero and sustainability strategies, the issue is bigger than digital transformation. Water and wastewater services are energy-intensive, and the relationship between water and energy is becoming more important as climate change, population growth and water stress increase pressure on both systems. The International Energy Agency notes that energy and water systems are deeply interdependent, with rising demand and climate constraints creating implications for both water security and energy security.

Digital tools can help utilities manage this pressure. Better data can support pump optimization, predictive maintenance, leakage reduction, wastewater process control and more efficient use of chemicals and energy. AI could also help utilities identify abnormal flows, forecast demand peaks, prioritise repairs and reduce unnecessary truck rolls by field teams.

But these benefits depend on reliable data pipelines. If operational data remains locked inside specialist systems, or if teams must manually re-key information between platforms, AI cannot provide timely or trusted insights. The Sustainability Matters article highlights this issue in the context of SCADA and control room systems, where some of the most valuable operational data exists but does not always flow cleanly into forecasting, reporting, or customer platforms.

The goal is not necessarily to replace specialist systems. In many cases, those systems are designed for critical operational tasks and should remain specialized. The key is interoperability: ensuring that data can move securely and consistently between operational, enterprise, and customer systems.

Regulation Raises the Stakes for AI Adoption

The governance challenge is especially important in Australia, where water and sewerage assets sit within the country’s critical infrastructure framework. The Australian Government’s Cyber and Infrastructure Security Centre states that the Security of Critical Infrastructure Act 2018 sets legal obligations for owners and operators of relevant critical infrastructure assets, including in the water and sewerage sector.

That means AI adoption in water utilities cannot be treated like experimentation in a low-risk commercial setting. Utilities must consider cyber risk, auditability, accountability, service continuity and public trust. An AI tool that recommends maintenance priorities, changes treatment settings or influences customer communications must be explainable and governed. Boards, regulators and customers will need confidence that automated systems are working from accurate data and that humans remain accountable for decisions.

This is one reason why the “move fast” model of technology adoption is poorly suited to essential infrastructure. Water utilities operate under obligations linked to public health, environmental compliance, billing integrity and continuity of service. If AI systems are introduced without clear governance, they may increase rather than reduce operational risk.

From Pilots to Enterprise Value

The next stage for many utilities may not be another AI pilot. It may be the less visible work of integration, data classification and governance.

This includes mapping where critical data sits, improving data quality, defining ownership, connecting operational and enterprise systems, strengthening cyber controls and establishing rules for how AI-generated insights can be used. It also means deciding which use cases are low-risk enough for early deployment and which require more robust oversight before automation is introduced.

The International Water Association has also highlighted the potential of digitalization and AI in water utilities, including the role of generative AI and agentic AI in supporting more advanced digital transformation. But the same direction of travel points back to the need for trusted, integrated information. Digitalisation turns fragmented datasets into usable intelligence only when utilities invest in the foundations first.

For utilities under pressure to modernize, reduce emissions and improve resilience, the practical message is clear. AI can help water companies make better decisions, but it cannot compensate for disconnected systems, weak governance or unreliable data. In fact, those weaknesses become more visible when AI is layered on top.

The utilities most likely to benefit from AI will be those that sequence the work carefully: build trusted data foundations, connect systems, prove value in controlled settings, then expand automation into higher-stakes decisions. That approach may be slower than headline-grabbing AI deployments, but it is more realistic for a sector where accountability cannot be delegated to an algorithm.

As water utilities move from being data-rich to decision-ready, AI will remain an important tool. The deciding factor will be whether the sector treats digital foundations as core infrastructure, not just an IT upgrade.

Source: www.sustainabilitymatters.net.au