Advancing AI adoption in Latin America and the Caribbean
1 - What is the LAC AI Acelerator?
The LAC AI Accelerator is a regional initiative that aims to foster the development and implementation of artificial intelligence (AI) use cases, especially agentic AI use cases, across Latin America and the Caribbean (LAC) countries.
The initiative will identify high‑potential use cases through an open call and support proponent teams in rigorously validating them. In doing so, the accelerator equips governments with evidence to decide whether to scale AI solutions with long‑term financing or to discontinue them. Lessons from the accelerator will be shared across the ecosystem to enable replication and scaling of effective AI use cases.
The initiative envisions empowered governments in LAC to harness AI to innovate in service delivery, promote a more efficient government, design data driven policy, and, ultimately, foster an inclusive digital economy.
The LAC AI Accelerator is the first iteration of what is expected to become an annual challenge led by the World Bank. Each iteration will be built on learnings from the previous cohorts and will include an increasingly more comprehensive toolbox of support mechanisms for the use cases.
2 - What is the value of participating in the LAC AI Accelerator?
The accelerator will deliver assistance to the proposing teams, called the Digital Service Teams, throughout the design and testing of the AI solutions.
The engagement will consist of the following:
Institutional maturity assessment: deep-dive evaluation of the proposing organization’s AI readiness, adoption and strategic positioning. This assessment will help the organization understand its current capabilities, identify gaps and define a roadmap for scaling AI initiatives effectively and responsibly.
Strategic advisory: expert guidance on selecting the most appropriate AI models.
Technical mentorship: hands-on companionship throughout the lifecycle of the pilot to troubleshoot issues as they arise and to test the use case. The World Bank Group and its strategic partners will not provide hardware nor software (including computing power, data and sandbox environments), therefore, all hardware and software will have to be put in by the Digital Service Team.
Specialized trainings to strengthen skills for AI-solutions deployment and to improve knowledge on AI fundamentals, data governance, and ethical principles to ensure safe implementation. Members of teams that participate in courses will receive a Certificate of Completion from the World Bank Group. Trainings will be tailored for policymakers and technical staff.
Networking and knowledge sharing opportunities. During the World Bank Digital Summit, the selected teams will be able to showcase their use cases, thereby expanding the visibility of their innovations and their opportunities for collaboration. The teams will have access to other types of World Bank Group online and in-person events.
3 - What are the direct beneficiaries of the LAC AI Accelerator?
The accelerator is open to teams from public or private sector entities, civil society organizations and Academia, all throughout the LAC region, interested in using AI to improve:
Category 1 – Digital government services (e.g., through AI agents with which citizens and businesses can interact and make transactions using natural language)
Category 2 – Public sector efficiency (e.g., through automated processing, simplification, and case management for Back-office optimization)
Category 3 – Policy design and delivery (e.g., through AI-assisted simulation of the potential impact of policy choices, such as subsidies, regulations or social programs, before implementation)
Applicant teams can use the Agentic State Vision Paper to guide the definition of their use cases.
Participating teams should be multi-disciplinary, with experts who understand both the technology and the specific development problem they are trying to solve.
It is recommended that proposals target one of the World Bank Group’s five priority sectors: infrastructure, agribusiness, health, manufacturing, and tourism. However, alignment with these sectors is not mandatory, and proposals from any domain that fit within the three eligible categories are welcome.
Participating teams must have the support of a public digital authority (national, state or local level) and will have to submit a commitment letter signed by the head of the relevant authority. Support may come from a consortium of public authorities within the same country or among different LAC countries; in this case, there will be more than one sponsoring organization and commitment letter.
4 - How will the Digital Service Teams be selected?
All Digital Service Teams will have to send the following documents to the address lacaiaccelerator_worldbank.org@worldbank.org:
The institutional commitment letter, that is, the letter signed by the head of the public digital authority that backs the Digital Service Team. Use the letter templates available on the right of this page.
The use case proposal document, that is, the record that describes the use case proposal, the problem it is trying to solve and the readiness of its AI solution. It should also include considerations for AI suitability and data, and a sustainability and scalability plan. Use the document templates available on the right of this page.
- The privacy notice through which the applicant consents to share personal information about himself/herself and the Digital Service Team members with the World Bank. Use the templates available on the right of this page.
Proposals will be evaluated by the LAC AI Accelerator Governance Committee, composed of a representative from each of the partner organizations, as well as experts from the World Bank Group’s Digital & AI VP.
5 - What are the next steps for the LAC AI Accelerator?
Here is the calendar of the regional initiative:
Milestone | Date |
Call for proposals | February 16, 2026 |
Reception of applications & eligibility screening | From February 16, 2026, to March 20, 2026 |
Announcement of selected use cases | March 30, 2026 |
On boarding to training | April 6, 2026 |
Monthly follow-up sessions | April 10, 2026 May 8, 2026 June 5, 2026 July 3, 2026 August 7, 2026 September 4, 2026 |
Closing ceremony | September 7, 2026 |
Publication of the Annual Report and use cases presentation | World Bank Digital Summit |
Knowledge sharing:
- The AI use cases will be stored in the AI Use Case and Knowledge Platform, developed through a collaboration among seven Multilateral Development Banks.
- The LAC AI Accelerator will produce an Annual Report featuring key lessons learned, best practices, and policy design and pilot evaluation recommendations. The report will serve as a regional reference and will be disseminated to policy makers and the international development community.
- Selected AI use cases will be showcased in regional and global forums, reinforcing the potential of AI to drive inclusive and sustainable development in LAC.
For any additional questions, interested parties can contact the organizing team by email at lacaiaccelerator_worldbank.org@worldbank.org.
6 – Frequently Asked Questions
Q1: What is the expected maturity level of the submitted use cases?
The LAC AI Accelerator is open to AI use cases at a wide range of maturity levels, from early-stage technology concepts to fully deployed systems operating under monitored conditions. What matters most is that the Digital Service Team provides a comprehensive description of the proposed solution and a candid assessment of organizational readiness, covering dimensions such as connectivity, compute capacity, digital public infrastructure, data availability, skills, and risk management. The Accelerator's support is calibrated to meet teams where they are.
Q2: The initiative emphasizes agentic AI, but the eligibility categories are broadly defined. Are purely non-agentic AI solutions eligible, or is there a minimum threshold of “agenticity” that proposals must meet?
The Accelerator prioritizes solving real development challenges with the most appropriate technology. While the initiative has a particular focus on agentic AI use cases, evaluation will place greater weight on the quality and clarity of the problem statement, the proposed solution, and the organization’s readiness (across policy, skills, infrastructure, and data) than on whether the solution is strictly agentic.
Q3: What specific criteria does the LAC AI Accelerator Governance Committee use to evaluate and rank proposals?
Proposals are assessed by the Governance Committee against five weighted criteria:
- Impact & Relevance: Does the solution address a significant development challenge and have the potential to reach a meaningful number of people?
- Data & AI Fit: Is AI, and specifically agentic AI, the right approach for this problem? Does the team demonstrate readiness in terms of policy, infrastructure, data, and skills?
- Feasibility & Team: Does the proposal present a realistic plan that a multidisciplinary team can execute within a six-month agile cycle?
- Safety & Responsibility: Does the solution protect data privacy, mitigate bias, and follow a “do no harm” approach?
- Sustainability & Scale: Is the solution designed to outlast the pilot phase and be replicable in other contexts or countries?
Q4: What type of public digital authority must back the AI use case?
Any public authority with competencies in digital matters is eligible to provide institutional backing, regardless of whether it operates at the municipal, state, or national level. The key requirement is that the authority issues a formal Institutional Commitment Letter, signed by its head, confirming political support for the use case and nominating a focal point to engage with the Accelerator throughout the process.
7 – Selected proposals
The LAC AI Accelerator supports 18 AI solutions across 12 countries in the region:
Country | Name of the AI solution | Description |
Colombia | Agente de IA para una Oferta de Servicio a la Ciudadanía Inteligente y Anticipativa para Bogotá | Built for the Alcaldía Mayor de Bogotá D.C. in Colombia, this AI agent is integrated into the district portal bogota.gov.co and connects public information with service delivery while orchestrating various key tasks within the citizen services system. |
Trinidad and Tobago | AI-Enabled Digital Public Services and Decision Intelligence Platform | Developed for Trinidad and Tobago’s Ministry of Public Administration and Artificial Intelligence, this AI-driven platform integrates predictive analytics and data tools to support government decision-making and strengthen public health system resilience and preparedness against climate-related disasters. |
Chile | Copiloto IA para recuperación automática de cupos liberados por cancelación en confirmación de horas | Deployed for the Municipalidad de Puente Alto in Chile, this AI co-pilot automatically detects cancelled or freed primary healthcare appointment slots and reassigns them to waiting patients using rule-based agentic automation, while providing dashboards to track recovered capacity and operational performance. |
Argentina | DINO EMPRENDE: Agente de Inteligencia Artificial para el Acompañamiento al Emprendimiento Femenino | Integrated into Chubut Province’s existing DINO bot in Argentina, this AI agent assists women entrepreneurs with personalized guidance on formalization, financing, training, and commercialization, while generating dashboards to support evidence-based public policy decisions. |
Chile | Ecosistema Integral de IA para Búsqueda, Cribado, Extracción, Síntesis y Consulta de Evidencia Clínica | To be deployed for the Ministerio de Salud of Chile, this AI-powered hybrid multi-model system automates the screening, extraction, and analysis of scientific literature (using NLP, RAG, and automated GRADE generation) to support technical health professionals in accelerating the biennial update of Clinical Practice Guidelines, reducing manual review burden while keeping human validation at the center of the process. |
Panama | Empleo 3.0: AI-Powered Public Workforce Development Platform | Integrated into an employment platform for Panama City’s Municipality, this AI system matches candidates to job postings, provides personalized career coaching with skills gap analysis, and offers labor market forecasting and program effectiveness tools to municipal decision-makers. |
Brazil | Fiscalização Inteligente de Contratos Complexos | Built for the Governo do Estado do Espírito Santo in Brazil, this AI system autonomously audits government contracts for regulatory compliance, generates risk alerts, and classifies invoices and deliverables by risk level to support contract oversight officials across the state executive branch. |
Brazil | Framework para Autoavaliação de Impacto Ético em Inteligência Artificial | Developed for Brazil’s Secretaria de Governo Digital, this AI governance tool guides federal public servants through a structured ethical self-assessment questionnaire to diagnose adherence to ethical principles and generate improvement recommendations for AI solutions developed or procured by federal agencies. |
Uruguay | Gobernanza de tecnologías emergentes: marco de valoración y sostenibilidad aplicado al Asistente Ciudadano | Implemented for the Intendencia de Montevideo in Uruguay, this decision intelligence process evaluates and pilots technological evolution options for the city’s citizen service virtual assistant, assessing agentic AI models using evidence on performance, privacy, sustainability, and public value. |
Peru | Iñatech: Tecnología de Soporte | Designed for local governments in the Department of Amazonas of Peru, this AI-powered predictive and prescriptive system (integrated into a mobile app) analyzes users’ mobility data to recommend and personalize ergonomic orthopedic devices made from Amazonian bamboo and 3D printing, with the objective of improving affordable access to mobility aids for elderly people and individuals with disabilities in rural communities. |
Mexico | Laboratorio UX/UI/AI en México para servicios digitales: Caso Chatbot “Xoli” | Created for the Agencia Digital de Innovación Pública of Mexico City, this LLM-based multilingual chatbot provides automated, geolocated information on cultural events, activities, and tourism services to residents and visitors. |
Turks and Caicos | National Health Insurance Board Agentic AI Treatment Abroad Program Case Manager Pilot | Designed for the National Health Insurance Board of the Turks and Caicos Islands, this agentic AI application orchestrates and manages end-to-end Treatment Abroad Program cases with human oversight. |
St. Vincent and the Grenadines | One Stop Vincy Gov AI Concierge | Deployed for the Government of St. Vincent and the Grenadines, this AI-powered concierge integrated into the national digital portal processes natural language queries to provide citizens with general information and personalized service status updates based on user identification. |
Costa Rica and the Dominican Republic | Operating System for Agentic Government | This AI solution is an open-source AI platform developed for the governments of Costa Rica and the Dominican Republic that enables citizens to access and complete public service procedures through a single natural language interface while coordinating workflows across government institutions. |
Colombia | Proyecto CEREBRO: Inteligencia Artificial para la Gobernanza Macroeconómica | Developed for the Alcaldía Municipal de Palmira in Colombia, this generative AI platform combines Colombia’s DSGE PATACON macroeconomic model with advanced AI to help governments simulate economic scenarios, evaluate fiscal and monetary policy impacts, and recommend optimal policy actions in real time. |
Uruguay | Resumen del Paciente | Developed for AGESIC and Uruguay’s Ministerio de Salud Pública, this AI solution automatically summarizes a patient’s full electronic health record, including chronic conditions, medications, allergies, and key procedures, to help physicians quickly access the most relevant clinical information within a standard 10-minute consultation. |
Argentina | Sistema de Inteligencia Artificial para la Optimización Predictiva del Servicio de Salud Municipal | Deployed for the Municipalidad de la Ciudad de Mendoza in Argentina, this AI-powered predictive system forecasts patient demand, appointment no-shows, and resource needs at municipal health centers to support operational staff in optimizing scheduling and medical resource allocation. |
Mexico | Ventanilla Digital: IA para la Eficiencia Gubernamental en Baja California | Built for the Agencia Digital del Estado de Baja California in Mexico, this AI-powered digital platform simplifies and digitalizes citizen-government interactions by reducing bureaucracy and improving public service efficiency through no-code technology, digital identity, and integrated e-government tools. |