Public trust is a critical factor in how effectively government agencies can serve their constituents — communication and engagement are key pillars of building trust. Today, with the help of generative AI tools, federal agencies can improve citizen experience with 24/7 access to resources across multiple languages.
That is a big value-add for integrating generative AI into government missions, according to Raj Rana, specialist and customer engineering leader for Google Public Sector. The ability to provide a conversational approach to search and virtual agents, whether in English or other languages, adds much to the overall government experience as it exists today.
“Generative AI, in conjunction with other technologies, is where you really get the biggest bang for your buck,” shared Rana in a recent interview produced by Scoop News Group, for FedScoop and underwritten by Google for Government.
He walked through a video demonstration of how a generative AI model might work along with a virtual agent to allow a constituent to ask questions about a tax return to the IRS. If the constituent was unable to get their questions answered, the virtual agent would connect them with a human agent to continue the conversation.
In the demonstration, Rana showed how a constituent who speaks Spanish could still interact with both the virtual agent and the human agent. The generative AI tool would facilitate direct translation in each instance.
“Looking at it from the perspective of the citizen, [this allows the government] to provide a holistic experience,” he added.
Rana explained that there are many areas where agencies can deploy generative AI. An ideal situation might be one that uses public-facing data to remove any additional challenges around the validation or classification of that data.
Additionally, Rana said, “Ensure that you’ve got the buy-in from subject matter expertise…It’s really, most valuable as an assistant to the existing government employee. [For example], how can we make them more effective? How can the technology work on their behalf?”
“And last, but not least, is to ensure that you have the alignment across all different groups. You bring in legal and procurement in and make sure that they’re aware of what this is and what this means so that as you’re deploying this, there won’t be any hiccups.”
For agency leaders interested in generative AI, it is critical to start with education, according to Rana, and learn what generative AI can and can’t be used for.
“This is a different model, even for folks who are very comfortable with traditional predictive AI,” said Rana, who said he has heard several conversations around the challenges of balancing the promise of generative AI with policy.
He recommended that leaders and “dreamers” in the organization use sandbox environments, like Google’s Discovery Zone — a solution offering for public sector customers — to validate and test out capabilities while still maintaining concrete boundaries that allow for conversations around policy and governance that are critically important.
Watch the full discussion with Rana and find out more about integrating generative AI into your agency’s mission. You can also hear from more stories from government leaders on Accelerating the Mission with AI and Security.
This interview was produced by Scoop News Group for FedScoop, and underwritten by Google for Government.