Adoption of Artificial Intelligence in Latin America and the Caribbean: Interest Surpasses the Global Average, with Geographic Concentration and Predominance of End-User Solutions

Last updated: 10/02/2025

    The rapid emergence of generative artificial intelligence and other AI-based technologies underscores the need for analytical tools that enable a better understanding of how these solutions are being adopted in the countries of Latin America and the Caribbean. Such tools are essential to provide evidence in support of the formulation and implementation of national AI strategies that foster productive, social and sustainable development.

    The methodology employs web traffic as a proxy for AI adoption, understood as the number of visits to websites offering AI solutions. An “AI solution” is defined as a product, service or functional system that integrates AI techniques as a core component to perform tasks requiring data processing, decision-making, content generation, personalization, recognition or prediction, either autonomously or semi-autonomously. Solutions that use AI only marginally or are not accessible to end-users are not included in the analysis. The sample (approximately 260 websites with the highest number of visits across 18 countries) was constructed on the basis of Similarweb and RankMyAI data, and is classified according to the functionality of the solution and its enabling technologies. It should be noted that the analysis does not capture the full breadth of the ecosystem. Solutions with limited web traffic, and in particular enterprise-level implementations in closed environments (back-end) — which are more typical of larger companies — remain outside the scope of this methodology.

Main Reults

1. A region with strong interest in the use of AI solutions. The region makes use of AI above what would be expected given its share of the global Internet user base. Worldwide, Latin America and the Caribbean account for 14 per cent of total visits to AI solutions, while the region represents only 11 per cent of global Internet users.

    

Distribution of visits and Internet users in Latin America and the Caribbean and worldwide, April 2025(Percentage of total visits to AI solutions and Internet users as a percentage of total population)

Source: own elaboration based on Similarweb.com and RankmyAI.com

2. Use of AI solutions concentrated in six countries.Web traffic data reveal a marked geographic concentration in six countries —Brazil, Mexico, Colombia, Peru, Argentina and Chile— which together account for 86 per cent of the regional total. This reflects the weight of their digital markets and the more advanced development of their innovation ecosystems. At a second tier, countries such as Ecuador (59 million), Costa Rica (20 million), the Dominican Republic (19 million), Venezuela (19 million), Bolivia (18 million) and Guatemala (17 million) register intermediate volumes, consistent with their population sizes and levels of digitization. Finally, in a third group, with between 6 and 15 million visits, are Panama, Uruguay, Paraguay, El Salvador, Honduras and Jamaica, whose lower absolute volumes are consistent with the scale of their markets.

3. Predominance of end-user AI solutions. With regard to types of AI solutions, in the region —as in the rest of the world— generative AI predominates, with a share slightly higher than the global average (78 per cent versus 74 per cent). By contrast, the adoption of more advanced solutions, such as open-source tools, development platforms and the use of AI models and APIs, is lower than the global average (22 per cent versus 26 per cent). The conclusion is clear: the region makes intensive use of end-user solutions but engages less in integration and production (less customization, reduced creation of local intellectual property, and lower participation in more complex value chains, among others). 

    

Visits to AI solutions in the world and in Latin America and the Caribbean, by enabling technology, April 2025 (Millions)

Source: own elaboration based on Similarweb.com and RankmyAI.com

    This trend is further confirmed when examining data by functionality, which shows that demand in Latin America and the Caribbean is concentrated in consumer-ready solutions with low technical requirements. Text generation leads with 69% of visits, followed by multimedia content with 19%. Business productivity solutions exceed 100 million visits, almost entirely oriented toward process automation, accounting for 6.5% of the total. Since large enterprises generally rely on internal platforms and private networks —not captured by this methodology—, it is plausible that most of this traffic originates from micro, small and medium-sized enterprises (MSMEs). The use of AI agents (4%) remains limited, suggesting an early stage of adoption.

Distribution of visits from Latin America and the Caribbean to AI solutions, by main functionality of the solution, April 2025 

Source: own elaboration based on SimilarWeb.com and RankMyAI.com

Analysis and interpretation

1. Why is Latin America and the Caribbean overrepresented in usage? The overrepresentation reflects a growing interest in exploring AI-based tools and reveals a dynamic potential for adoption. First, it is driven by the high penetration of mobile Internet (70% of the population in 2024) and smartphones (81%) (GSMA, 2025), which enables broad access to cloud-based applications. Second, the region’s young and digitally active population rapidly adapts to the use of new technologies. Third, the simplicity of generative AI interfaces allows interaction through a web browser or mobile application without requiring advanced hardware or technical training, facilitating use by individuals with limited digital skills as well as by MSMEs and self-employed workers.
2. Why is usage concentrated in six countries? The leading countries combine higher levels of connectivity, stronger entrepreneurial capacity, and a broader local and regional supply of services. In addition, sectoral policies and predictable regulatory frameworks reduce adoption frictions (e.g., personal data protection, public procurement, risk guidelines), increasing the propensity to experiment. This concentration may also reflect network effects surrounding platforms and technical communities.
3. Why does end-user consumption predominate? The limited weight of open-source development and the use of models or APIs suggests bottlenecks in: (i) specialized talent (e.g., data engineering, machine learning, security, model governance); (ii) data infrastructure and access to cloud services for experimentation and deployment; (iii) sophisticated business demand capable of driving integration projects; and (iv) coordination among academia, the public sector, and technology firms to transform pilot initiatives into scalable solutions. Without addressing these constraints through public policies—particularly those aimed at productive development—the region risks consolidating itself as a net consumer of AI, with limited capacity to adapt solutions to local contexts and a growing dependence on external providers. 

Conclusion

    Latin America and the Caribbean (LAC) is overrepresented in global AI usage, driven by the traction of end-user generative solutions and their accessibility through web and mobile platforms. However, this adoption is concentrated in a handful of countries and reveals an incomplete maturity: the region consumes much but produces and integrates little. The public policy priority—linking digitalization with productive development—must be to transform the intensive use of AI into productive and innovative capacities. This entails strengthening productive linkages, fostering diversification, scaling, and local and regional value chains, and integrating MSMEs into these chains. To achieve this, it is essential to close gaps in talent, infrastructure, and data governance, and to deploy digital public goods and pro-competitiveness instruments—such as standards and interoperability, open-source development, innovative public procurement as a driver of demand, and regulatory sandboxes—that can move the region from plug-and-play adoption toward programmatic integration of AI into processes and value chains. The ultimate goal is to generate local, interoperable solutions with tangible impact on productivity, formalization, sustainability, and social and territorial inclusion.

Prepared by: Valeria Jordan and Laura Poveda.