The OECD.AI Policy Navigator

Our policy navigator is a living repository from more than 80 jurisdictions and organisations. Use the filters to browse initiatives and find what you are looking for.

Taina (Public Service Agent)


Added by:   OECD analyst
Added on:   03 Aug 2026
Updated by:   OECD analyst
Updated on:   03 Aug 2026

Taína is a public service agent developed by the Government Office of ICT (OGTIC) to help citizens access information about public services and procedures in clear and accessible language. It provides guided assistance through a web interface. The initiative explores how public AI can improve citizen service while remaining ethical, inclusive, and open source.

Name in original language

Taina (Agente para Servicios Públicos)

Initiative overview

Taína is a public artificial intelligence initiative developed by the Government Office of Information and Communication Technologies (OGTIC) of the Dominican Republic, in collaboration with Universidad del Caribe (UNICARIBE) and the Center for Public Intelligence (CPI), an applied research lab focused on public AI systems. The initiative was launched as a research and development (R&D) project to explore how artificial intelligence could be applied to citizen assistance in an ethical, inclusive, and context-aware manner. CPI contributed to the conceptual and technical design of the public AI infrastructure, while UNICARIBE supported academic validation and testing environments. Taína provides guided assistance covering all services available in the official public service catalogue of the Dominican State. Through natural language interaction, citizens can consult requirements, procedures, documentation, timelines, and institutional responsibilities in clear and accessible language. The system structures and interprets catalogue information to make it easier for people to navigate government services. In its current version, Taína operates through a web interface and has been validated in academic and citizen environments to test usability, clarity, and accessibility. The next phase of the initiative includes integration with the Government Call Center, enabling citizens to interact with Taína via telephone as well as digital channels. The objective is to embed the AI agent across the State’s omnichannel service system, ensuring consistent responses through web, phone, and future service touchpoints. Taína is designed as public AI infrastructure rather than as a proprietary product. Built under open-source principles, it enables collaborative development and oversight by public institutions, developers, and universities. Guided by principles of ethics, inclusion, accountability, and public value, the project seeks to position artificial intelligence as a shared digital public good that strengthens access to information and public services across the Dominican Republic.

Other relevant details

Results, outcomes and impacts: Taína currently provides structured information on more than 800 public services across 21 government institutions. The system was validated through five structured user testing sessions in academic and controlled citizen environments. Participants reported an average satisfaction score of 4.7/5 when consulting public service information. Testing assessed clarity, usability, and comprehension of procedures. Future phases include integration with the Government Contact Center IVR system to enable AI-assisted phone responses and expand omnichannel citizen support. Challenges and lessons learned: The initiative faced challenges related to unstructured and dispersed government data, limited in-house technical AI capacity, and uncertainty around compute requirements for scaling. Public procurement frameworks require detailed cost specifications in advance, which is difficult for AI systems whose infrastructure needs evolve during deployment. Rigid purchasing schemes also slowed experimentation. These challenges were addressed through research partnerships with academia and technical collaboration agreements with major technology providers, enabling controlled testing and flexible compute access. A key lesson is that public AI requires adaptive procurement models, strong data governance, and institutional capacity building. Successful implementation depends on political support, cross-institutional coordination, and updated regulatory frameworks that allow iterative development while maintaining accountability.

About the policy initiative


Category:

  • AI policy initiatives, programmes and projects

Initiative type:

  • AI use cases/projects in the public sector

Status:

  • Active

Start Year:

  • 2024

Target Sectors:


OECD AI Principles: