Smarter cities, safer communities: How state and local government leaders are advancing public services with AI

State and local governments are moving beyond pilots and proofs of concept to real-world applications — and they are doing this with agents. We are now firmly in the agentic era, where 55% of public sector leaders say their organizations have already adopted AI agents in production, according to the Google Public Sector ROI of AI in the Public Sector report. Furthermore, 42% of these organizations have already deployed more than 10 distinct agents, and 46% report that their workforce productivity has at least doubled thanks to these tools. Respondents also anticipate allocating more than 50% of their future AI budgets to agents.  

“Leaders across industries and around the world are harnessing this new technology effectively and at scale to solve real problems and to drive value and impact,” said Elizabeth Moon, Managing Director of Customer Engineering, Public Sector Sales at Google Public Sector. “The ROI we're talking about today is that ability to deliver faster, more personalized and equitable services and outcomes.”

Across state and local government functions, including transportation, public safety, day-to-day service delivery and planning for major special events, modern public administrators are tapping into AI to build responsive and data-driven experiences. Google Cloud Next convened forward-thinking state and local leaders from across the country to share how their organizations are re-imagining burdensome legacy technologies and frameworks in favor of AI-powered, automated and unified platforms.

From reactive road safety to automated, predictive insights

The transportation sector’s approach to road safety has traditionally been reactive; the safety rating of an intersection, for example, might be based on how many fatalities and crashes have already occurred there. According to the National Highway Traffic Safety Administration, there are a staggering 40,000 fatalities due to traffic crashes annually. Transportation leaders are working to materially reduce this number — and they are leveraging modern AI and data solutions to inform their approach.

“Our number one priority is what we’re doing to reduce serious injuries and fatalities in our transportation system,” said Toks Omishakin, Secretary of the California State Transportation Agency (CALSTA). To that end, CALSTA partnered with Google Cloud and Deloitte to create the Roadway Safety Insights (RSI) tool.

Previously, integrating dozens of records “to analyze all the data related to how to solve a particular safety challenge,” Omishakin said, “would take weeks and months.” By combining AI simulations, machine learning models and data, CALSTA is now able to condense that time to minutes to draft reports and even detect where crashes are likely to happen before they take place. 

“The RSI tools … are more and more focused on analyzing all that data and using it to say, ‘here are some predictive outcomes we can come up with for roads that haven’t even had a crash yet,” Omishakin explained.

These AI-driven efficiencies apply to more than safety — AI and automation can be further leveraged in transportation to speed up analysis, generate cost estimates and link them to project prioritization and more.

“At the end of the day, transportation is still an engineering problem,” said Anant Dinamani, Principal at Deloitte Consulting LLP. “We’re also using AI to speed up transportation demand planning at the local levels. What took 2.5 years to build — a transportation demand plan — they’re doing it in months.”

A data-driven approach brings efficiency and equity to city challenges

As the State of California brings proactivity to road safety, the City of Los Angeles is focusing on its MyLA311 system. The city processes nearly 3 million service requests and inquiries through 311 annually. Among the most common are reports of illegal dumping or graffiti, street pavement issues and bulky item pickups.

“Reimagining is no longer having people tell me there’s a problem, which is really what 311 is, instead we just know there’s a problem and — crazy thought — fix the problem before everyone has to tell us,” said Ted Ross, CIO for the City of Los Angeles. Ross views AI as the connective tissue across the city’s enterprise, bridging disparate data and entrenched legacy processes. By moving beyond personal productivity to automated business workflows, this architecture transforms previously fragmented tasks into a single, intelligent flow that multiplies his staff's capabilities.

Imagine technologies like computer vision augmenting pothole identification capabilities and sending an automated robot to fill the pothole in the middle of the night. That is the type of proactivity Ross is envisioning for more efficient city management — one that Tim Kelly, Mayor of the City of Chattanooga, echoed.

“We have our Center for Urban Informatics that we’ve integrated, we’re bringing all that data into BigQuery, not only trying to get to Vision Zero and eliminate pedestrian fatalities,” Kelly said, “but also anticipate where we can improve traffic flow or put new crosswalks in by watching what people are doing and then using AI to really optimize.”

Aside from bringing greater efficiency to finding and prioritizing traffic and transportation issues across Chattanooga, Kelly adds that automated solutions would also make city services more equitable. Too often, he noted, the same people show up to city council meetings, and those with the “loudest voices” end up driving public policy. 

“We’re not there yet, but we’re close to a model where we can ingest a lot more data on a regular basis to give us lights to steer by so we know we’re actually on the right path,” Kelly said. 

Rebuilding legacy systems into a single source of truth

Before modern applications and AI agents can automate workflows, however, state and local technology leaders need a strong, unified foundation to support a sophisticated AI stack. Creating that foundation with a complex tangle of legacy systems and siloed data, however, is no small feat. Carlos Braceras, Executive Director of the Utah Department of Transportation, referred to the organization’s previous method of modernization as “Frankensteining” legacy systems piece by piece. 

To break this cycle, UDOT bypassed traditional silos by deploying Google Vision AI to complete a massive statewide property parcel identification task. A project originally estimated to require 33.5 years of manual labor was completed in just under 12 months, proving that AI can help compress decades of work into days.

“As we wanted features or new capabilities, we kept adding piece by piece. Now we’d created this system that is so disjointed with data everywhere. It's a mess and it’s impossible to take advantage of the new technologies,” Braceras said. “This is where you have to say, ‘I’m just going to push that aside,’ or ‘we’re willing to start from scratch.’”

“Transportation is at an inflection point where there’s lots of good data available. There are a lot of great systems that are all built in individual silos that are all serving needs for different parts of the engineering organizations,” Dinamani said. “But the future is about integration, understanding how safety is affected by construction, how work zones affect maintenance, how all of that affects operations.” 

Preparing the workforce for AI transformation

In addition to a solid technology foundation, AI success also hinges upon a transformation of culture and skillsets. Some of the biggest barriers to adoption can be psychological, rooted in the fear that automation will displace human workers. “Helping to encourage the idea that these are tools to supplement our work, not replace our teams,” is critical to moving forward, Omishakin said. 

Moreover, these advanced tools are only as effective as the people using them. As Ross stated, “If I gave you a Ferrari and I took a proficient race car driver and gave them a Honda Accord, they would embarrass you … the person behind the wheel makes more difference than the wheel itself.”

Investing in human capital is as fundamental as investing in the tools they will be using, a leadership priority Kelly stated plainly: “There are two types of mayors in the United States: the ones who realize workforce development is the most important work we’re doing, and the ones who will soon realize it.”

For the City of Los Angeles, the AI revolution is a more expansive opportunity to train staff not only in AI skills but in comprehensive digital skills. 

According to Ross, we have to use this opportunity to upskill agency talent. “You’re not just training them in how to run a prompt, you’re training them in the next wave of innovation. If you can capture them now, and build those skillsets now, they will be prepared and positioned for whatever is next.”

Ready to accelerate your mission impact in the agentic era? Tune into the Gemini for Government: The Blueprint for Mission Impact webinar for a strategic deep dive into building proactive, automated public services at scale.