In its first week live, an artificial intelligence agent called Bobbi resolved 82% of citizen queries across three police forces in the United Kingdom without once passing the call to a human.
That kind of result is what UiPath and Google were selling to governments this week.
UiPath announced in a Tuesday (May 5) press release agentic AI capabilities inside its Automation Suite, built specifically for government agencies and regulated industries that cannot send sensitive data to the cloud.
The update lets agencies run AI models inside their own infrastructure, the release said. That means choosing between cloud-hosted models from Anthropic, Google and OpenAI, or fully self-hosted models that never leave the agency’s environment. The platform runs across Amazon Web Services (AWS), Microsoft Azure and OpenShift, so agencies can use the infrastructure they already have. Data sovereignty controls and built-in compliance tooling are included from the start.
“As government agencies adopt agentic automation, our focus is helping them stay in control of their data, models, and how AI is used,” UiPath Public Sector Chief Technology Officer Chris Radich said in the release.
The practical point is that agencies waiting on the sidelines because of data concerns now have a procurement-ready path forward. UiPath is not asking government buyers to change how they operate. It is bringing agentic AI to where they already are.
Advertisement: Scroll to Continue
Google Goes to Work for Government
Google Cloud Next ’26, made a similar statement at a much larger scale by announcing new tools, Let’s Data Science reported Tuesday. The Gemini Enterprise Agent Platform is designed to connect data, people and operational goals across large institutions. Gemma 4, Google’s latest open model, gives agencies a deployable option they can customize on their own terms. New eighth-generation TPUs provide the compute infrastructure for agents running at volume. Google Cloud reported its models now process 16 billion tokens per minute through its direct API, up from 10 billion in the prior quarter.
Google’s choice of use cases told the larger story. Transportation, healthcare research and campus operations all featured as examples. That framing positions agentic AI not as a productivity tool but as mission-critical government infrastructure. For public sector buyers still evaluating their options, the signal from Cloud Next was that the largest cloud providers have decided this market is a priority.
Government at App Speed
Both product announcements land as public expectations run ahead of government capability. Salesforce Government Cloud President Kendall Collins wrote in January that citizens now expect from their government the same speed they get from a banking app. Every slow touchpoint and redundant form signals to citizens that government is falling behind. That gap does not just frustrate people. It erodes trust.
The piece pointed to governments that are already closing it. Singapore’s “Ask Jamie” virtual assistant has fielded more than 15 million queries across 80 government websites, Collins wrote. In Barcelona, a centralized platform gives civil servants a 360-degree view of every citizen interaction across departments, turning disconnected processes into coordinated, real-time service. Bobbi, the U.K. police agent, is part of the same pattern.
Estonia is running the most ambitious version, piloting a network of interoperable agents that cross agency lines so that a citizen renewing a passport works with one system that coordinates everything on their behalf, Collins wrote. A Salesforce survey found that 90% of global respondents are willing to use an AI agent to interact with their government.
For all PYMNTS AI coverage, subscribe to the daily AI Newsletter.