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.

VitrineGov – Brazil’s Public Sector Innovation Platform


Added by:   OECD analyst
Added on:   16 Nov 2026
Updated by:   OECD analyst
Updated on:   16 Nov 2026

VitrineGov is Brazil’s public platform for finding, assessing, documenting and scaling proven government practices. Its AI engine, Veridia, supports pre-selection, evidence checking, correlation analysis, interview preparation and audit trails, turning tacit know-how into explicit, reusable one-page assets. The result is faster learning, safer curation and greater replication across government.

Name in original language

VitrineGov – Plataforma de Inovação do Setor Público Brasileiro

Initiative overview

VitrineGov is a public-sector innovation platform led by Brazil’s Ministry of Management and Innovation in Public Services (MGI). It was created to solve a structural government problem: valuable innovations are often developed inside one agency, but remain fragmented, weakly documented and hard for other institutions to understand, trust and reuse. VitrineGov addresses that gap through a full curation cycle that prospects initiatives, screens submissions, conducts pre-selection, interviews teams, documents implementation journeys, publishes accessible one-page public summaries and monitors continuity, results and replication potential.

Artificial intelligence is embedded in this workflow through Veridia, the AI engine used by the curatorial team. Rather than replacing public servants or making binding decisions automatically, Veridia strengthens expert review. It helps to analyse initiative submissions considering VitrineGov’s five criteria — creativity, relevance, results, adaptability and continuity — while organising evidence, testing the correlation between public problem and solution, preparing the semi-structured interview, and producing auditable outputs such as evidence comparisons based on interview transcripts, complementary risk matrices and capability-building pathways.

What makes VitrineGov distinctive is that AI is not treated as an isolated chatbot or productivity shortcut. It is part of a governed public methodology inspired by knowledge-conversion logic: tacit knowledge from teams and programmes is converted into explicit, reusable public assets. Each published initiative becomes a one-page, one-click reference for peer learning, visibility, mentoring and adaptation. The platform therefore delivers not only recognition, but also a practical transfer mechanism that lowers the cost of replication across agencies.

VitrineGov is already active on the Ministry’s website and connected to the National Programme for Management and Innovation (PNGI), which gives it national scale. Its next phase is to deepen analytics, expand monitoring, strengthen auditability and widen federative reuse across more organisations, states and municipalities. In the long term, the ambition is to make public innovation searchable, comparable, trustworthy and transferable at scale — so governments spend less time reinventing solutions and more time adapting what already works.

Other relevant details

Results, outcomes and impacts: VitrineGov is already operational and currently showcases around 30 federal and state initiatives. Its scale is amplified by PNGI, which ended 2025 with the participation of 22 states and 16 municipalities, creating a national field for reuse and adaptation. Published cases are being used to give visibility to proven practices, support peer exchange and turn selected initiatives into national references. Internal data reported by the team indicates that published initiative pages have already exceeded 10,000 visits. Challenges and lessons learned: The main challenges are to compare heterogeneous initiatives fairly, avoid overclaiming when evidence is uneven, preserve accountability when AI is introduced, and adapt promising practices across very different institutional contexts. VitrineGov responds with strong human oversight, explicit public-interest criteria, staged review, semi-structured interviews with initiative teams, transcript-based evidence checking, and an auditable trail linking claims to documentation. A second lesson is that replication requires more than publishing a success story: it requires structured documentation, focal points, governance routines, risk mapping and capability development. For that reason, the model couples curation with complementary maturity plans, including risk matrices and training pathways to close implementation gaps. The broader lesson is that governments gain more value from AI when it is embedded in a trusted institutional workflow, not deployed as a stand-alone tool.

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:

  • 2022

OECD AI Principles:

—