The city of Tucson, Ariz., has announced its first major agentic AI platform investment — specifically, an automated support agent — as outlined in a recently opened request for proposals (RFP).
Tucson’s City Council received public pushback in 2025 related to a proposed data center development project in the area, a subject that has been dubbed “the NIMBY issue of our time.” AI, however, offers potential efficiencies for local governments. Chatbots are one critical tool that can help open the front door to services.
Now, Tucson is “seeking proposals from qualified vendors to deliver an AI-powered service automation solution,” per its RFP.
Tucson’s RFP marks the city’s “first foray into agents and the potential for future agentic solutions,” according to an email from the city’s CIO Chris Mazzarella. It aims to help address a high turnover in call center roles; other local governments, too, have turned to AI to address call center staffing challenges.
“We hope this will augment our call center teams, help in future training for new hires and free staff up to handle higher complexity tasks, while enabling any savings to be re-applied to other operational improvements or taxpayer/customer savings,” Mazzarella said.
Tucson’s AI agent will do two things. First, it will function as a typical large language model chatbot to provide basic information. But it will also act as an agent with the means to process requests, update records, and perform like a staffer, the CIO said.
Rather than being trained only on publicly accessible data, the agent will also be trained on the city’s business knowledge base and practices, and will be connected to Tucson business tools and systems of record.
An AI agent for the city of Tucson will need the same level of oversight and training as a human staff agent — and in the early training period, even more so, the CIO said.
The city opted to work with an external vendor for this initiative rather than building it internally, Mazzarella said. That’s because developing such an AI agent is the primary business of some companies, and residents expect a high level of functionality and security from the initial launch.
“I’m banking on external experts being able to continually innovate, maintain, secure and ensure effective AI services beyond what we could do with a small team and investment internally,” Mazzarella said, noting that once the agent is built, city officials will then be tasked with guiding business operations and product management aspects.
The city will use a phased implementation method, starting with the IT Department service desk and Tucson Water Customer Service Solution Center. The city’s AI Program Manager Kate Carter will help guide the solution’s internal testing within the city’s IT help desk before applying it to Tucson Water, which has previously leveraged tech from VODA.ai for its infrastructure systems.
Tucson Water’s use of the new solution should yield the first direct benefits to residents, the CIO said, as customers are slated to receive improved access to services, information in their preferred language, and access to a human staff member when needed. A phased implementation helps create a “foundation,” the CIO said, on which the city can implement across departments. RFP submissions are due by July 23.
This AI implementation, Mazzarella said, will help the city assess operational readiness, cultural needs, training and governance in order to more effectively consider other opportunities for AI-enabled automation.
Julia Edinger is a senior staff writer for Government Technology. She has a bachelor’s degree in English from the University of Toledo and has since worked in publishing and media. She’s currently located in Ohio.
See More Stories by Julia Edinger