Abstract
Background Whereas it took ChatGPT less than a week to achieve a global base of one million users, many of us were not counted among the early adopters. As an industry, the water utility sector tends to be a slow follower when it comes to advanced technologies. Consequently, this reluctance to organize and take advantage of new opportunities like Generative Artificial Intelligence (AI) can compromise operating efficiency, cost management, and service delivery. Why does this happen? What needs to change? How can water utilities become just as receptive and quick to adopt advanced technologies as other industries? The answers to these questions point to organizational change management. It involves the people-side of change, how change begins with preparing the organization for something different; how employees need to be convinced of better, more effective ways than current methods to accomplish desired results. This presentation contextualizes change management practices within the realm of ChatGPT. It is intended to stimulate and accelerate the adoption of transformative technologies in the water utility sector. What follows is an authentic and realistic story on the adoption of ChatGPT at a recognizable water utility known for innovation-DC Water. DC Water's Approach DC Water's awareness of ChatGPT began shortly after its public release (version 3.5) in November 2022. Exploratory efforts commenced following the launch of ChatGPT 4 in March 2023. (The anticipation of ChatGPT 5 and Microsoft's Bing Chat Enterprise release before UMC 24 further underscores this technology's rapid advancement.) As will unfold in this presentation, DC Water realized enough potential of ChatGPT's value to justify introducing it to the organization for further evaluation. In early May 2023, DC Water's Innovation Director conducted a 'Lead and Learn' training session to approximately 18% of the total workforce. Soon afterwards forty-five employees requested to learn more about ChatGPT in promoting its use at DC Water. A pivotal aspect to adopting ChatGPT was the establishment of a Community of Practice (CoP). As a diverse group of curious employees, it was tasked with crafting a vision and establishing five key objectives to set expectations. Six S.M.A.R.T. goals were set in place to target and time-bound specific measurements. These goals addressed operating efficiency, staff engagement, employee satisfaction, decision making, industry leadership, and knowledge dissemination. To facilitate ChatGPT's adoption, a comprehensive 27-page 'QuickStart' User Guide was curated. This enabled employees to harness ChatGPT's full potential governed by a set of best practices, including guidance on: (1) its ethical use, privacy needs, and reliability verification; (2) definition and examples of ChatGPT command usage; and (3) actual use case examples of it being employed at DC Water. What proved to be most challenging was scaling ChatGPT to higher rates of employee adoption. Many employees struggled in going from traditional methods to leveraging ChatGPT to research industry data, extract and analyze information, and make decisions. The CoP members played a crucial role here, donning the hats of both educators and change agents. Benefits and Significance To demonstrate this need, DC Water conducted periodic reviews and outcome assessments to validate ChatGPT's tangible impact on operating efficiency, research capabilities, and decision making. More than just the potential to benefit, DC Water had already proven its ability to generate efficiencies, strengthen research, and make timely and better decisions by what got accomplished. As of August 2023, CoP members identified thirteen different applications of ChatGPT for consideration. Examples included AI-enabled customer support, IT coding assistance, industry research (problem solving), financial analysis, records retrieval for Freedom of Information Act (FOIA) support, and action plan generation. It is anticipated that other ChatGPT applications will emerge as more is learned of its potential to contribute to ongoing DC Water needs. A salient point of contention is the misconception of ChatGPT as a replacement for people. While there are tangible benefits, like with most tools, ChatGPT still requires human intervention to generate commands, evaluate results, and make the right decisions. It can liberate employees from routine, time-consuming work to focus elsewhere on higher value areas. Going beyond routine tasks to more sophisticated use cases, DC Water experienced how Engineers (problem solvers) can leverage ChatGPT's large language model (LLM) to acquire high-quality data for decision support, saving time and money verses starting fresh without industry and public data readily available. Another point of contention is how a transformational innovation like ChatGPT can lead to large-scale automation threats. Social media is full of dystopian predictions that many believe are overblown. While utility leaders should acknowledge this concern, DC Water was able to strike a balance between tool utility and risk mitigation. DC Water will share its version of the 'dos and don'ts' of ChatGPT usage for further guidance. From an intangible selling point, employees felt more productive when using ChatGPT. They experienced a feeling of empowerment in executing commands, interpreting results, and making decisions. PoC members were motivated to experiment and use ChatGPT for purposeful outcomes. Preliminary Results Together, the tangible and intangible benefits experienced so far, while impressive, are still being tabulated. As of August 2023, preliminary results in using ChatGPT show time savings, cost efficiencies, and employee enthusiasm. Feedback indicates a positive impact on employee attitudes and behaviors based on a non-random sample size of forty-five staff (3.7% of the total headcount). Whether CoPs or other approaches are employed to influence the people-side of change, the desire is to create a ripple effect across the entire organization. To go from slow to quick follower, utility leaders need to formulate and act on a change management strategy. They must address the practical challenges of adopting advanced technology, and how this can lead to a 'demand pull' from others wanting access to ChatGPT capabilities. Shared at the UMC conference will be discussion on using a CoP approach to implement ChatGPT throughout the organization. DC Water will provide detailed metrics on operating efficiency and employee satisfaction. The aim is to give conference attendees a sense of what can potentially be realized and how when embracing change. DC Water will describe how it engaged employees in group activities and one-on-one assistance for specific need fulfillment. This proved to be particularly helpful in keeping its newly-formed CoP motivated to learn and promote ChatGPT among work peers. While this case study centers on DC Water, the strategies and lessons gleaned are universally applicable, offering other utilities a blueprint for technological integration and organizational transformation. Conclusion DC Water's presentation offers invaluable insights into the nuances of organizational change management, specifically within the context of ChatGPT's integration into a water utility. Much of its success was due to innovation leadership and a CoP approach among ChatGPT enthusiasts from across the organization. Actual data on DC Water's efforts to adopt and scale ChatGPT across the organization will be presented for conference attendees to consider, including benefits, lessons learned, and governing factors to counter risk. As it looks ahead, DC Water envisions further deepening its collaboration with AI tools, exploring avenues like predictive maintenance and asset management. Its journey with ChatGPT is just the beginning of a broader tech-driven transformation. DC Water believes that advanced technologies like ChatGPT offer the power and potential to revolutionize how work is done within the water sector and beyond.
This paper was presented at the WEF/AWWA Utility Management Conference, February 13-16, 2024.
Author(s)R. Bornhofen1
Author affiliation(s)DC Water 1;
SourceProceedings of the Water Environment Federation
Document typeConference Paper
Print publication date Feb 2024
DOI10.2175/193864718825159241
Volume / Issue
Content sourceUtility Management Conference
Word count7