From data to deployment: Advancing responsible use of AI agents in US state government

High-quality data is the foundation of trustworthy and effective AI deployment since AI systems depend on accurate, well-managed data to generate reliable insights. Although nearly 9 in 10 CIOs view data quality as critical, fewer than 1 in 4 have formal data quality programs in place,⁴ making poor data governance a persistent challenge for many states. Inaccurate or incomplete data can be amplified by AI, leading to misinformed decisions, resource misallocation, and reduced public trust. 

Improving data management requires consistent organization, validation, and secure sharing across agencies. These practices promote transparency and accountability while ensuring that AI-generated insights are accurate and fair. Investments in clean, well-structured data can serve as the blueprint for every responsible AI application that follows. 

From data management to scaling AI agentsOnce strong data governance is in place, the next step is scaling AI responsibly across operations. Three factors are essential to this success: interoperability, security, and open standards.

1. Interoperability ensures that systems can share and interpret data across agencies, improving coordination, reducing administrative burdens, and minimizing errors.

2. Security protects sensitive data from misuse, aligns with compliance standards, and sustains public trust.

3. Open standards make AI systems easier to audit and adapt, supporting transparency, collaboration, and long-term innovation.

Tools, such as Elastic Agent Builder, enable governments to apply these principles in practice. Agencies can use Agent Builder to create custom agents that collaborate securely across systems using open standards. For instance, agencies can automate case management, enabling different departments like the Department of Health and Human Services, Department of Labor, and Department of Education to share data and coordinate workflows while maintaining role-based access controls and compliance. Built on a foundation of reliable, well-governed data, Agent Builder can help government workers generate insights, automate tasks, and transform constituent services.

No state can advance AI responsibly in isolation. Public-private collaboration provides the expertise, technical frameworks, and accountability structures for safe and effective AI adoption. Private developers contribute engineering depth and awareness of emerging risks while government leaders ensure that deployments align with policies, ethics, and public interest.

Partnership models can include co-developing interoperability frameworks, conducting pilot programs in controlled environments, or establishing shared governance structures to strengthen oversight and data quality. Through collaboration, states can accelerate innovation while maintaining security, transparency, and public trust. 

Governments that commit to principles of clean data, open architecture, and active collaboration will not only modernize their operations, but also demonstrate what trustworthy, user-focused AI looks like in practice. Elastic stands ready to support states in turning that vision into a working model that others will follow.