In a significant shift, government agencies are beginning to embrace artificial intelligence as a tool for enhancing service efficiency. A new AI agent named Bobbi has showcased remarkable efficacy in its first week by resolving 82% of citizen inquiries across three police forces in the UK without needing human intervention. This development underscores a growing trend where public sector bodies are looking to AI to bridge service gaps while maintaining data integrity.
Emergence of Agentic AI in Government
leading in this movement, UiPath has introduced agentic AI capabilities within its Automation Suite, specifically tailored for government and regulated industries. This suite allows agencies to deploy AI models on-premises, providing a significant advantage for organisations that handle sensitive data which cannot enter the cloud. By leveraging existing infrastructures such as Amazon Web Services (AWS), Microsoft Azure, and OpenShift, agencies can adopt AI solutions without overhauling their operational frameworks.
Chris Radich, Chief Technology Officer for UiPath Public Sector, emphasized the importance of data control, stating, “As government agencies adopt agentic automation, our focus is helping them stay in control of their data, models, and how AI is used.” This approach addresses the hesitance of agencies that have been waiting to implement AI solutions due to concerns over data security and compliance.
Google's Strategic Positioning
In parallel, Google has made strides in establishing its AI offerings for the public sector. At the Google Cloud Next ’26 event, the company unveiled the Gemini Enterprise Agent Platform, intended to enhance operational efficiency within large institutions. This platform is designed to connect data with personnel and organizational objectives, thus positioning agentic AI as essential infrastructure rather than a mere productivity tool.

The release of Gemma 4, Google’s latest open model, presents agencies with customizable deployment options. With the introduction of eighth-generation Tensor Processing Units (TPUs), Google has significantly enhanced computational capabilities, claiming that its models now process an impressive 16 billion tokens per minute—up from 10 billion in the previous quarter.
Meeting Public Expectations
This dual push from UiPath and Google comes at a critical time when public expectations are evolving. Citizens increasingly demand the same responsiveness from their government agencies as they experience with modern banking applications. The slow pace of traditional bureaucratic processes can lead to frustration and a decline in public trust. As Salesforce Government Cloud President Kendall Collins noted earlier this year, “Every slow touchpoint and redundant form signals to citizens that government is falling behind.”
By integrating AI agents into their operations, government agencies can not only improve service delivery but also restore faith in public institutions. The solutions provided by UiPath and Google represent a procurement-ready path for agencies eager to modernize their operations while ensuring compliance and security.
Implications for Future Government Operations
The strategic initiatives by these tech giants signal a turning point for government agencies, suggesting a broader acceptance of AI as a necessary tool for enhancing public service efficiency. As more agencies adopt these technologies, the potential for AI to reshape government operations will become increasingly apparent. The focus will likely remain on balancing data sovereignty and the need for agile, responsive service delivery.
The future landscape of government services will likely be defined by how effectively these AI solutions can be implemented while addressing public expectations and operational realities. As more agencies engage with these technologies, the evolution of public service delivery will provide a fascinating case study in the intersection of technology and governance.
The stories that move AI & crypto markets — before the market reacts.
Free. 7am ET. Five stories. 62,400 readers.



