Artificial intelligence won’t replace water operators anytime soon. But it may help utilities deploy scarce capital more effectively, surface buried institutional knowledge, and make operations more adaptable.
That was the consensus at a well-attended panel on AI in the water sector, held during WEFTEC 2025 and hosted by the newly launched Water-AI Nexus Center of Excellence.
Moderated by Radhika Fox, former Assistant Administrator for Water at the U.S. EPA, the session featured leaders from the LA Department of Water and Power (LADWP), Houston Public Works, and Xylem.
“We’re gathering here at a really pivotal moment for the water sector,” said Fox. “AI is not a silver bullet, but another tool in our toolbox.”
Data first, then AI
Before AI can deliver operational value, panelists stressed the need for foundational investments in cloud migration, data governance, and executive accountability.
“We have created a foundation where AI is not a use case. It’s part of the foundation of how we’re going to build the business,” said Janisse Quiñones, CEO of LADWP.
This includes hiring a chief AI officer, organizing IT, business, and process teams into cross-functional ‘three-in-a-box’ groups (where each initiative is jointly led by representatives from operations, IT, and process improvement), and cleansing datasets to ensure model readiness.
Houston City Engineer O.J. McFoy underscored the importance of data control: “We guard it with our lives. That’s step one — keeping the data in house.”
Operational use cases take root
Utilities shared real pilots underway today:
- Knowledge capture: LADWP is using AI to build chatbots trained on interviews with retiring staff, letting new hires ask questions like, “How do you operate valve #123?”
- Process optimization: Houston is using AI to fine-tune chemical dosing and process controls at treatment plants and to support a $4 billion plant upgrade with digital twin modeling.
- Leak detection: Houston lost 32 billion gallons to leaks last year. McFoy called that “unacceptable” and said AI is being deployed to detect and localize leaks using sensor data and machine learning, helping reduce non-revenue water by pinpointing losses and optimizing response.
Quiñones also flagged a future use case that could have industry-wide implications: AI-assisted capital planning.
For example, “If you invest a dollar here, your return is better here and your risk is reduced,” she said.
Culture shift: From fear to fluency
Panelists agreed that workforce buy-in is essential, especially in union environments.
“We let anybody that wants to have Copilot have Copilot,” said Quiñones, referring to Microsoft’s AI assistant. “[We] ensure they understand it’s Copilot, not pilot.”
LADWP also created a grassroots AI community of practice. “Our teams are teaching each other,” she said. “And it’s pushing our union leadership to sit at the table.”
With union members starting to adopt AI tools, leadership wants to help shape how these technologies are integrated into their work.
McFoy emphasized the opportunity to attract younger, tech-savvy staff: “We’re putting in new leaders. They are ready to take our utilities, take our cities, to the next level.”
Rewiring the enterprise
Xylem CEO Matthew Pine urged utilities to think bigger than pilots.
“Use cases are good, but it’s really about reconstructing how you do business,” he said. “AI can help you do that in a pretty dramatic way.”
Pine said utilities should think of AI as the next Six Sigma, a transformational discipline requiring internal champions.
Xylem has trained 16 “AI champions” who support adoption across the company, plus full-time “AI navigators” who help business units scope and execute projects.
To succeed, Pine said, culture must shift. His three pillars: “Inspire to innovate. Empower to lead. Accountable to deliver.”
Looking ahead
Asked what will change in five years, McFoy pointed to lower losses and smarter capital plans: “We will not be losing 32 billion gallons … We will have two brand new plants in which we are keeping pace with the growth of Houston.”
Quiñones projected that LADWP will meet its goal of 70% local water with AI helping track risk, schedule, and cost: “AI has been a part of the fundamental fiber of who we are as a utility.”
Pine predicted widespread role changes: “80% of jobs … are going to be completely different in two years.”
Fox closed with a broader hope: that AI could finally help the sector communicate its value.
“What if AI helped us finally crack that nut?” she asked. “That we’re not out of sight, we’re not out of mind — but people understand the value of the utilities today.”
