Water is essential for life. For thousands of years, human settlement and advancement has been dictated by a reliable supply of clean, safe water. In the face of a fickle supply system, people flourished, moved or perished. While the need to effectively and efficiently manage water has not changed over time, the management tools available to the market most certainly have. Water utilities are now equipped with data-driven technologies that allow them to tease out previously-unattainable information about water supply and demand, empowering water managers to build a dependable clean water future.
One of the newest tools for assisting in smart water management is software powered by artificial intelligence (AI). The first software decision support systems used hard-programmed “expert systems” or “rule-based algorithms” to derive outputs or decisions by incrementally narrowing the available choices. While functional, these systems rapidly become unwieldy in the face of dynamic conditions and competing interests. The new tools of AI are changing that. AI uses pattern recognition applications that employ a set of inputs, weighting factors, summation and transfer functions that, by definition, are autonomously and dynamically updated as new information is presented.
In many ways, AI programming replicates the way humans learn. During the “learning phase” of AI programming, input data is correlated to known outputs to allow the algorithms to learn over time. Then, in the “operational phase,” the program begins to make sense of patterns as new data is introduced. Because of AI’s ability to constantly adapt and process large amounts of data in real-time, it is an ideal tool for managing water resources in an ever-changing environment, and the business of water, allowing water utility managers to maximize current revenue and effectively plan for the years ahead.
Utilities today have two clear choices: seek new supply to deal with resource volatility, or optimize existing use within the available supply. By using AI and new software-as-a-service platforms, as well as low cost sensors and affordable communication networks, managers can create dynamic strategic, tactical and financial operations for their utilities. Municipalities can also better plan and execute capital project requirements, better understand real-time water loss, more efficiently operate distribution networks and ensure maximum revenue capture — all while meeting ever-increasing customer demands.
The power of AI unleashes the imagination of our water professionals. For example, AI-driven planning can combine growth projections with future trends in water availability and infrastructure condition assessment to maximize the impact of investment in infrastructure. AI can also be used to automate the inspection of sewer systems or the accuracy of meters.
The benefits go beyond the utility. AI-enabled online platforms offer customers a personalized, engaging and informative way to view real-time water consumption, pay bills and access information about dynamic water resource conditions. These online portals also use big data to educate consumers about conservation and nudge them along a path where small changes in behavior add up to substantial cost savings for them while enabling large-scale water resource savings for the community.
AI, however, is only as good as its data and the understanding of the output. For optimal results, AI requires the availability of an adequately sized, validated data set which includes the information that can characterize the problem, and train and test the network. Equally important is a fundamental understanding of the problem. While AI can handle massive data sets and computations, it is not failsafe. How many of us have been directed down a back lane or country path following an “optimized” GPS route?
AI is not, then, a fire-and-forget solution. A key feature of a successful application is a necessary detailed understanding of the environment to both ask the right questions and evaluate the applicability of the response. Critically interpreting the outcome becomes increasingly important as algorithms perform more of the grunt work and basic analyses, ultimately abstracting the problem and depriving humans of some of that insight. Interpreting the output is going to be a fundamentally important job requirement for utility professionals. Machines will most certainly perform this work more quickly, with fewer errors and with greater precision, but human interpretation is still needed — at least for now. Questioning and investigating — which are uniquely human traits — will remain important and knowledgeable water utility staff will continue to be a critical and essential element of delivering safe, reliable water to utility customers.
To access the value delivered by AI systems, water utilities need to create the processes around which data is collected, accessed and analyzed. This requires the creation of an ecosystem in which there is unfettered access to primary data and other supporting data sets, as well as a means of normalizing and storing this data combined with a team of skilled professionals to make sense of it all.
AI will fundamentally change the way water utilities operate and manage water resources in an increasingly volatile environment. The water sector’s immediate challenge is to structure data and smart water management services to maximize AI’s potential. Once this is addressed, the industry can then focus on developing dynamic models that run continuously to optimize utility operations, increase customer delight and best protect our shared clean water resources.
As the cost associated with AI declines and adoption becomes more widespread, machines and humans will collaborate to pull these water use observations into actionable information for use in the years ahead. Our task today is to start.