top of page

Leveraging AI with Amanda Willis and Gigi Karmous-Edwards

Our theme article features the wisdom provided by our two speakers during our October 15 event, “Leveraging Artificial Intelligence: Smart Solutions for Institutional Knowledge Retention in Water Utilities.” Gigi Karmous-Edwards, President, Karmous Edwards Consulting and Amanda Willis, Senior Water and Wastewater Specialist, The Water Tower, explored effective use of artificial intelligence (AI). Our blog post covers their overview of AI, exciting use cases and reports, and important takeaways and next steps you might take in incorporating AI into your work. 


The Great Lakes PRESERVE team met with Gigi Karmous-Edwards, President of Karmous Edwards Consulting, and Amanda Willis, Senior Water and Wastewater Specialist with The Water Tower, on October 15th at 11 AM EST to discuss “Leveraging AI: Smart Solutions for Institutional Knowledge Retention in Water Utilities.” 


ree

Overview of AI 


Karmous-Edwards began with a fundamental question: How can we bridge the divide between resources and technological capabilities?   From there she walked us through broad AI concepts, such as the difference between generative AI, machine learning AI, and deep learning. We also discussed the importance of harnessing unstructured data, which can be defined as images, audio files, spreadsheets, or emails. Karmous-Edwards cited that 80% of a utility's data is unstructured data. For that reason, AI can open up the possibilities to offer data-driven insights for these data in a much more accessible way.  


Ethical Considerations 


Karmous-Edwards recommends members purchase enterprise level LLMs and turn OFF data sharing in order to allow for more ethical experimentation. The speaker team also summarized an exciting recently released report (which Moonshot participated in), titled: GenAI as a Catalyst for Water Sector Transformation.



The report found GenAI to be practical and affordable, running no more than $60/user/month. The paper also found AI tools to be useful on day one, including data analytics on structured and unstructured data. The report also encourages utilities to start thinking about data governance on unstructured data to offer guidance to utility employees.  


Use Cases 


Our next speaker, Amanda Willis, offered practical takeaways and insights that utilities might consider when taking the next steps in AI. After recognizing the careful and important work that operators do each day, she offered caution when giving AI data, including having trust systems in place.  


Willis demonstrated her team’s use of AI through a tool called “What Would Jerry Do?” In this initiative, she and Karmous-Edwards worked with a team to create a list of questions. They interviewed Jerry, an 82 year old operator who had been working for the city of Carlton, GA for almost 40 years. They were able to input Jerry’s answers into a GPT that now allows existing operators to ask the GPT, a quantifiable means of retaining institutional knowledge, before asking Jerry himself. Willis’s experience has also opened up possibilities for capturing rate increases, retaining institutional knowledge, and tapping into the many possibilities that AI can represent. 

 

Final takeaways from our speakers include:  


  • Don’t be intimidated: An in-depth knowledge of code is not necessary to make these ideas a reality. Not everyone can afford a cloud, or have the knowledge, but open-access enterprise solutions such as these GPTs can give operators the chance to test their systems, see how it works, and continue to add possibilities.  

  • Experiment and Explore: Both speakers encouraged members to purchase an enterprise level LLM, turn off data sharing, and explore the possibilities that it presents with pdf files, emails, and even asking it to demonstrate what it is capable of. Start by thinking about pain points to discover the possibilities of AI. 

  • Lean in: It's important to view AI a co-pilot, not an author, and a tool that can increase our capacity and create new career pathways such as Digital Water Engineer or Digital Water Technician.  

  • Consider Workplace and Culture: Develop a CoP (such as PRESERVE) around experimenting with AI, and ensure that you have buy-in from the decision makers in your company.

  • Build Trust Networks: AI development is is a liminal space now. Innovations are happening faster than research can be published.  Take time and intention to build trust networks that are engaging with AI to stay current with functionality and security when using water system data. 

     

Both speakers expressed a sincere amount of gratitude and enthusiasm, and we are certainly so grateful for their insights! Please reach out to us at PRESERVE@moonshotmissions.org if you have questions for our speakers or would like to discuss the possibility of AI in your work.  


To align with this month’s theme, we’ve also compiled some resources on AI we hope you’ll find helpful! 



Please email us if you've found additional resources and tools useful. We'd love to continue the conversation!

bottom of page