From Pipes to Predictive Analytics: American Water Charts AI's Course

COLUMBUS, OH – June 16, 2026 – In a conference room filled not with Silicon Valley developers but with state utility commissioners and industry executives, the conversation has turned to artificial intelligence. Here at the Mid-Atlantic Conference of Regulatory Utilities Commissioners (MACRUC), the presence of a panel titled "Navigating AI" signals a profound shift underway in one of America's most traditional and essential sectors: water.

Leading the discussion is American Water, the nation's largest regulated water utility. With its Chief Technology & Innovation Officer, Deb Degillio, taking the stage, the 140-year-old company is positioning itself at the forefront of an operational revolution. The move underscores a critical transition for an industry built on iron pipes and concrete reservoirs—a pivot from a purely physical infrastructure business to one increasingly driven by data, algorithms, and predictive analytics.

The Digital Deluge: AI in Modern Water Management

For decades, the operational model for water utilities has been largely reactive. A pipe breaks, crews are dispatched. A water quality issue is detected, treatment processes are manually adjusted. American Water, along with other forward-looking utilities, is now leveraging AI to move from this reactive stance to a predictive one.

The foundation for this shift is data. The company's rollout of Advanced Metering Infrastructure (AMI), or smart meters, provides a torrent of real-time information on water usage. This data does more than just automate billing; it allows the utility to detect potential leaks on customer property and identify anomalies in the distribution system. Paired with machine learning and Geographic Information System (GIS) mapping software, this data helps predict which sections of its vast network of pipes are most likely to fail, allowing for proactive maintenance and targeted capital investment.

"American Water remains committed to leveraging technologies that enhance customer satisfaction, while building resilient systems and delivering essential services to our customers and communities every day," said Cheryl Norton, the company's EVP and Chief Operating Officer, in a statement affirming this strategy.

This commitment extends beyond leak detection. Across the sector, AI is being explored to optimize complex water treatment processes, ensuring water quality while minimizing chemical and energy use. Algorithms can forecast demand spikes during heatwaves or predict contaminant loads after heavy rainfall, enabling plants to adjust operations preemptively. In an era of increasing regulatory pressure around contaminants like PFAS, the ability to automate monitoring and reporting through AI is becoming less of a luxury and more of a necessity.

Regulating the Algorithm

The very presence of this conversation at a regulatory conference like MACRUC highlights the central challenge: as utilities adopt AI, how do regulators ensure it is used safely, fairly, and transparently? The 'Navigating AI' panel is just one of several sessions here dedicated to the topic, including a critical discussion on the "Ethical and Governance Challenges Associated with AI in Utility Regulation."

Regulators are moving from a world of auditing physical assets and financial records to one where they must scrutinize algorithms. A key concern is the "black box" problem, where complex AI models make decisions that are difficult for human operators to interpret or explain. "Regulators aren't just asking 'does it work?'; they're asking 'can we trust it, can we audit it, and is it fair?'" noted one industry analyst. This is crucial when an AI system is adjusting chemical doses for a public water supply.

In response, a new regulatory landscape is taking shape. According to recent industry surveys, 43 states now require pre-approval for AI systems that automatically adjust treatment processes. Furthermore, with water systems designated as critical infrastructure by the Cybersecurity and Infrastructure Security Agency (CISA), any AI deployment must come with ironclad cybersecurity protocols to protect against manipulation or attack. The dialogue here in Columbus is focused on building frameworks that ensure AI-driven decisions are consistent with due process, equity, and the public good.

A 140-Year Pivot: From Infrastructure to Information

For American Water, celebrating its 140th anniversary this year, the embrace of AI represents the latest chapter in a long history of evolution. Founded in 1886, the company built its reputation on the reliability of its physical network. Today, it is clear that maintaining that reliability for its 14 million customers depends as much on information as it does on infrastructure.

This is the "operational innovation" that often goes unseen by the public. Customers won't see the algorithm that predicts a main break, but they will experience its benefit: uninterrupted service. They may not understand the machine learning model analyzing their water usage, but they will appreciate the proactive alert about a costly leak in their home.

By placing a technology and innovation chief on a prominent stage before its primary regulators, American Water is making a strategic statement. It is signaling that technology is no longer an ancillary IT function but a core component of its business model, essential for managing billions of dollars in assets and fulfilling its mandate to provide safe and affordable water. This pivot is a blueprint for how legacy industries can reinvent themselves to meet 21st-century challenges.

The Unseen Costs and Future Currents

Despite the immense promise, the path to an AI-powered water sector is not without obstacles. Many utilities, particularly smaller ones, are hampered by legacy systems, a lack of high-quality data, and budget constraints. There is also a significant need for workforce development to equip employees with the skills to work alongside these new digital tools.

More fundamentally, the industry is beginning to grapple with a profound paradox: the technology meant to help manage water is itself incredibly thirsty. The massive data centers that power AI require vast amounts of water for cooling. A recent United Nations report highlighted this growing concern, with some projections estimating that global water consumption for AI infrastructure could reach 6.6 billion cubic meters by 2027.

"The industry must balance the operational gains from AI against the technology's own environmental and resource costs," cautioned an environmental policy consultant. "It's a critical part of the sustainability equation." This creates a complex challenge for utilities and regulators alike, forcing them to weigh the efficiency gains of predictive analytics against the direct resource consumption of the underlying technology. For an industry built on managing a finite resource, the true test of this technological pivot will be whether its digital gains outweigh its very real-world thirst.