City staff outlines AI rollout for utility call center and other services; supervisors to pilot call summarization and coach features

City IT staff on Wednesday described a multi-stage plan to deploy artificial intelligence tools in the utility call center and other city services, saying the work aims to reduce hold times, free up staff time for higher-value tasks and improve security monitoring.

Ronnie (IT staff), who led the presentation, said the city already uses some machine-learning features on its network-perimeter security tools and is expanding AI to customer care. "One of the first things we're going to implement... is call summarization," Ronnie said, explaining that the system would transcribe calls and present draft notes to call takers to review, saving several minutes per call.

Planned features: staff listed several capabilities planned in stages through the fall and year-end: automated call transcription and summarization for faster documentation; a supervisor "whisper mode" that alerts and allows a supervisor to privately coach a call taker during live calls; chat and email queues for the utilities department monitored and routed with AI; a call-taker chatbot the reps can query in real time for process guidance; and continued expansion of AI-based cybersecurity monitoring that can take autonomous defensive actions. Ronnie said the system will require training from historical calls and policy documents and that staff will stage deployment to ensure accuracy and safety.

Council members raised questions about safety and unexpected behavior from large language models. When a council member asked if the city had observed aggressive behavior from AI in other tests, Ronnie referenced public cases involving other vendors and said the city is using closed models and has not experienced those problems in production. Staff also discussed possible future integration with next-generation 9-1-1 systems and whether some after-hours call-taking might be returned from outsourced vendors to city-managed AI-assisted workflows.

Why it matters: The city is seeking operational gains in customer service and efficiency while balancing privacy, security and accuracy. Staff emphasized staged rollout, human review and supervisory oversight; no regulatory or policy changes were voted at the meeting.

Implementation and schedule: Staff said the work requires ingesting historical calls, policies and other data for training and that some features are expected before October with additional capabilities by the end of the year. The IT team will report back to council and bring features online incrementally.