AI is being adopted across Latin America to boost prediction and automate business decisions, delivering efficiency and cost gains where data, training and governance are in place; however, most evidence is descriptive, geographically concentrated and highlights persistent challenges around data quality, transparency and resistance to change.

Main Finding

AI is emerging as a transformative but context-dependent tool for business decision-making across Latin America. When implemented with appropriate training, governance, and strategic alignment, AI improves predictive capacity, automates processes, and supports data-driven decisions—delivering efficiency gains, cost reductions, and opportunities for strategic innovation. However, these benefits are uneven because of data-quality limits, resistance to change, transparency concerns, and algorithmic bias.

Key Points

• Study scope and outcome
  • Systematic review following PRISMA 2020; final sample: 27 studies selected from an initial 276,302 Scopus records (2021–2025).
  • Selected articles were original, open-access, conducted in Latin American countries, and published in Spanish or Portuguese.
• Geographic concentration
  • Most studies originate from Peru, Colombia, Chile, and Ecuador, indicating research concentration in a subset of countries.
• Sectoral coverage
  • AI applications reported across industry, agriculture, finance, education, and public services.
• Primary uses of AI
  • Enhancing predictive capacity, automating processes, and supporting data-driven decision-making.
• Benefits observed
  • Increased operational efficiency, reduced costs, and facilitation of strategic innovation.
• Key enablers for effectiveness
  • Staff training and human capital, ethical governance frameworks, and alignment with organizational strategy.
• Persistent challenges
  • Organizational resistance to change, limited or poor-quality data, concerns over transparency and interpretability, and risk of algorithmic bias.

Data & Methods

• Review framework: PRISMA 2020 guidelines for systematic reviews.
• Database and timeframe: Scopus search covering publications from 2021 to 2025.
• Search approach: Boolean search terms combining AI-related and decision-making–related keywords.
• Inclusion criteria:
  • Original research articles
  • Open-access
  • Empirical focus on Latin American countries
  • Published in Spanish or Portuguese
• Screening filters applied:
  • Temporality (2021–2025), geographic focus, language, document type, accessibility (open access), duplication removal, and relevance screening.
• Final sample: 27 studies retained for synthesis out of 276,302 initial hits.
• Country distribution: predominantly Peru, Colombia, Chile, and Ecuador (other Latin American countries underrepresented).

Implications for AI Economics

• Productivity and growth
  • AI adoption can raise firm-level productivity and generate cost savings, but gains depend on complementary investments (human capital, data infrastructure).
• Adoption heterogeneity and returns
  • Economic returns to AI are likely heterogeneous across sectors and firms; small firms and those in data-poor environments may realize smaller gains without targeted support.
• Labor markets and skills
  • Effectiveness of AI in decision-making depends on staff training; there are implications for skill-biased labor demand, retraining needs, and potential short-term displacement.
• Inequality and regional divergence
  • Concentration of research and deployments in a few countries suggests potential widening of digital and economic divides across the region unless diffusion accelerates.
• Market structure and competition
  • AI-driven efficiency and automation could change competitive dynamics (scale advantages for data-rich firms); regulation may be needed to ensure fair competition.
• Governance, transparency, and trust
  • Algorithmic bias and opacity undermine economic benefits by raising legal, reputational, and adoption costs; governance frameworks that promote transparency and accountability can increase adoption and social welfare.
• Policy levers
  • Policies that support data quality, public data sharing, workforce reskilling, SME adoption subsidies, and ethical standards will likely increase the social returns to AI.
• Research gaps relevant to economics
  • Need for causal and quantitative studies on AI’s impacts on productivity, wages, employment composition, firm profitability, and inequality in Latin America.
  • More geographically diverse and longitudinal evidence is required to assess long-term economic effects and spillovers.

Suggested next steps for researchers and policymakers: prioritize causal impact evaluations, invest in data infrastructure and training programs, develop governance standards for transparency and fairness, and create policies to support diffusion of AI benefits to SMEs and underrepresented countries.