Building Agentic AI in the Public Sector

For most organizations, deploying AI is about how fast you can ship. Public institutions face a different bar: governance, transparency, and explainability have to be design constraints from day one.

Spain's Agencia Estatal de Administración Digital (SGAD) drives digital transformation and manages shared technology across national, regional, and local governments. Working under strict data protection rules and on-premises infrastructure, the team has built a suite of citizen-facing assistants, and is now moving from classic chatbots to transactional, agentic AI that serves both constituents and regulators.

In this fireside conversation, José Carlos Martínez Durillo, Head of the Innovative Services Department at SGAD, sits down with Alan Nichol, CTO & Co-founder of Rasa, to share what it actually takes to design responsible agentic AI for public institutions at scale and what the future holds for AI in government.

What we'll cover:

• What responsible agentic design looks like when every decision has to be legible to regulators and citizens.
• The trade-offs public sector teams actually hit when they ship and how to navigate them without compromising trust or capability.
• How SGAD is evolving from conversational assistants to WhatsApp chatbots that complete real transactions (documents, certifications) across departments
• Lessons on piloting, staged rollout, and scaling AI under strict data-protection and infrastructure constraints that any government team can borrow.