AI is accelerating a new era of public sector innovation

The advancement of generative AI marks a technology paradigm shift impacting every part of the public sector. A newly released study Google commissioned with GovExec highlights the pervasiveness of this technology as 94% of respondents reported they plan to use AI in their department or agency in the immediate future. Brent Mitchell, Vice President of Go-to-Market at Google Public Sector said, “We stand at this pivotal moment in time. This is a new era of American innovation where AI is reshaping every aspect of our daily lives.” 

Ushering in a new era of innovation

While a majority of state and local government respondents to the newly released study Google commissioned with GovExec believe AI will improve overall efficiency and productivity, some of the biggest concerns surround skilled staff. The solution? Combine the deep policy expertise of public sector analysts and staff with the technical expertise of industry partners. In fact, 98% of respondents indicated they will rely on external resources to implement AI.

Strategic partnerships and collaboration offer state and local governments a path to greater efficiency, streamlining access to citizen services and providing personalized engagements by tapping into the incredible strides industry has already made in AI and other emerging technologies. It’s not just about creating cutting-edge technology, it’s about accelerating the adoption of those technologies in a way that benefits the nation.

For Google, “we are really focused on building a stronger, safer, more competitive America in this AI era — that is our mission,” Mitchell said. “We are going to deliver mission outcomes, but we are going to do it with a unique full-stack approach to AI.”

This approach requires advancing not just AI itself, but all of the foundational layers supporting it. From data centers and chips, to developer training and productivity, to research and model development, to the platform itself, every part of the ecosystem surrounding AI solutions will be impacted by secure, integrated solutions enabling accelerated innovation. Such an all-encompassing approach drives efficiency at every level. 

Taking pilots to mission-critical deployment

Several recent use cases illustrate the impact AI is already having on the public sector landscape. Covered California, partnering with Google and Deloitte, leveraged Document AI to automatically validate 84% of documents in real-time. This enables the organization to verify 50,000 documents per month. 

The partnership also enabled Covered California to streamline security and compliance with Assured Workloads, a Google Cloud solution that automatically supports FedRAMP compliance, and Google Security Operations, which scans all log information, signatures and threats to eliminate security blind spots and proactively safeguard against attacks.  

In Missouri, through partnerships with Google Public Sector and Quantiphi, the Missouri Department of Social Services implemented virtual agents and saw a 70% reduction in average wait time. The virtual agents can respond to 32,000 requests per month, and only about half elect to move forward and speak with a live agent. 

Vertex AI Platform is particularly critical to optimizing generative AI development and accelerating AI application development and delivery. Vertex AI is a fully managed and unified platform for developing and deploying secure AI agents, applications and models to rapidly deliver solutions at scale. Using Vertex AI, public sector organizations can streamline operations, empower government workers and, ultimately, accelerate service delivery to the benefit of the community.

Vertex AI provides access to Google’s latest Gemini models, now natively multimodal, and Gemma open models. With Gemini, it’s possible to process up to 2 billion tokens in a single product. This translates to reasoning more than 3,000 pages of text in one product. 

“You're trying to navigate a pile of paperwork, or, say, a stack of invoices — you can do it with no special prompting,” said Jason Gelman, Director of Product Management for Vertex Generative AI at Google Cloud. “You can also process two hours of video with audio at the same time.”

So many state and local government functions require processing and analyzing seemingly endless amounts of paperwork. With the help of AI, these processes can become significantly faster and more efficient, giving time back to human agents to devote to higher-level tasks. 

Consider disability claims, for example, Gelman said. Using Gen AI capabilities, “for agency workers, we're helping shorten the time to process claims, and for citizens, we're helping submit claims and get your benefits faster.”

New York State’s AI-powered technology transformation

Ahead of the Gen AI Live & Labs event, Google Public Sector and the State of New York announced a first-of-its-kind partnership through which New York State employees will have access to advanced technology tools to help them enhance disaster recovery, service delivery and overall digital transformation. At the event, Dru Rai, Chief Information Office of New York State, shared his vision for the outcomes of this partnership. 

“At the end of the day, we serve New Yorkers and our goal is very simple: [enhancing] quality of service, speed of service and cost of service,” Rai said. “They tell me you can’t have all three at the same time, but that’s what we try to do.”

Successfully implementing AI requires “peeling the onion back,” optimizing layer by layer, he said, mirroring Google’s full-stack approach. Employees will have access to transformative AI tools, but “first you have to train employees, then you have to train the model, then you peel the next onion, then those models have to be trained in context.”   

Government regulations and compliance needs bring added complexity to this process. For example, in New York, data from certain agencies cannot, by law, be shared with other agencies. So how can government technologists share learnings and maximize efficiencies while contending with legally required silos?

By “creating private models which can do all these things, and yet understanding the limitations of these models and making sure that … we don't cross the line when it comes to applying the law,” Rai said. 

As Rai touched on various use cases — processing hundreds of thousands of transactions ingested per minute, enhancing customer experience, implementing human-centered design and modernizing systems — he noted that, “in almost every scenario, Gen AI can help. It won’t be perfect, but it will be a great starting point.”

The key to modernization, he added, is not simply rewriting what was written on mainframe, but thinking differently, changing the whole business process, meeting citizens where they are and offering the right omnichannel experience to help them get to where they need to go. 

While AI opens up many possibilities, it’s not a panacea. AI implementation needs to be thoughtful and planned-out. It’s not enough to have a “fancy model,” Rai said, without data quality that’s up to par, specific outcomes to work toward and the right solutions to meet goals.

“Start with the goal, then apply the model, train the model, pull the data, fix the data, train your people, and then hopefully you get the goal,” Rai advised. “This tool is not going to give you a goal, this is a ‘how.’ You still need to define your goal, you still need your use cases, and the human beings are going to drive these tools.”  

This is the first in a series of articles based on Google Public Sector Gen AI Live & Labs, in collaboration with GovExec, which convened industry experts and leaders across city and state governments, as well as higher education, to discuss this new era of innovation. To help keep you at the forefront of the latest advancements, sign up today to receive the Google Public Sector Newsletter.