AI can materially raise productivity and enable new services for SMEs in developing economies, but adoption in places like Botswana is held back by skills shortages, poor infrastructure, financing gaps and weak data governance. Policymakers should pair phased, firm-level capability building with ecosystem financing and pragmatic data rules to make benefits broad and inclusive.
Artificial Intelligence Adoption for Sustainable Development in SMEs: Challenges, Strategic Pathways, and Lessons for Developing Economies
A structured literature review finds that AI can materially boost SMEs' efficiency, personalization, and innovation in developing-country contexts (illustrated by Botswana), but real uptake is constrained by skill gaps, infrastructure, finance, and fragmented data governance.
This study critically examines the adoption of artificial intelligence (AI) by small and medium-sized enterprises (SMEs) in developing economies, using Botswana as a primary context. Employing a structured narrative literature review, this study synthesizes existing empirical evidence, theoretical frameworks, and policy analyses to elucidate the multifaceted enablers and barriers influencing AI integration within resource-constrained SME environments. Drawing on the Resource-Based View (RBV), Dynamic Capabilities (DC), Technology-Organization-Environment (TOE), and Diffusion of Innovation (DOI) frameworks, this study highlights AI’s role in enhancing operational efficiency, customer personalization, and innovation capacity, thereby fostering competitive advantage and contributing to sustainable development. Key enabling factors include supportive government policies, innovation ecosystems, financial access and external market pressures. Critical constraints include skill shortages, infrastructural deficits, financial challenges, and fragmented data governance. This analysis advocates for strategic, phased AI adoption approaches, emphasizing incremental capability development, ecosystem collaboration, targeted human capital investment, infrastructure enhancement, scalable financing models, and ultimately, sustainable outcomes. The policy implications underscore the need for adaptive regulatory frameworks, inclusive financial mechanisms, and robust data governance to cultivate an environment that enables inclusive and sustainable AI adoption. This integrated analysis offers actionable insights for policymakers, practitioners, and SME stakeholders to accelerate inclusive and sustainable digital transformation and economic diversification in developing contexts.
Summary
Main Finding
AI offers significant value-creation potential for SMEs in developing economies (illustrated with Botswana) — improving operational efficiency, customer personalization, and innovation capacity — but widescale, inclusive adoption is constrained by skills shortages, infrastructure gaps, limited finance, and fragmented data governance. The study argues for strategic, phased adoption supported by ecosystem collaboration, targeted human-capital investment, improved infrastructure, scalable financing models, and adaptive regulation to achieve sustainable and inclusive digital transformation.
Key Points
- Contribution: Integrates RBV (Resource-Based View), Dynamic Capabilities (DC), TOE (Technology–Organization–Environment) and Diffusion of Innovation (DOI) frameworks to explain AI adoption dynamics in resource-constrained SMEs.
- AI-driven value creation:
- Operational efficiency: automation, predictive analytics, inventory/supply-chain improvements.
- Customer experience: personalization via analytics, chatbots, marketing automation.
- Decision making & innovation: real-time analytics for scenario planning, risk management, business-model adaptation.
- Main enablers: supportive government policy (e.g., Vision 2036, SmartBots), innovation hubs (BDIH), external market pressures, growing mobile broadband penetration.
- Main barriers: skill shortages, poor organizational readiness, unreliable digital infrastructure, limited access to affordable finance, fragmented data governance and regulatory uncertainty.
- Strategic pathways recommended: phased/incremental adoption (from simple automation → analytics → advanced AI), ecosystem collaboration (public–private partnerships, hubs), targeted training/upskilling, infrastructure investments, tailored financing instruments (e.g., scalable credit, blended finance), and robust data governance frameworks.
- Policy recommendations: adaptive and inclusive regulatory frameworks, public investment in digital infrastructure and human capital, targeted SME financing, and clear data governance to foster trust and interoperability.
- Contextual fact: Botswana (2024) — household internet access 68.9% (mobile broadband 86.1%); individual internet use 57.5% — indicating foundational digital readiness but organizational capacity gaps.
- Limitations: structured narrative review (not a systematic/meta-analytic study); dependent on published literature/policy documents and thus constrained by the availability of firm-level empirical studies in developing contexts.
Data & Methods
- Method: structured narrative literature review (chosen to synthesize fragmented empirical, theoretical, and policy sources where systematic reviews may be premature).
- Search strategy:
- Databases: Scopus, Web of Science, Google Scholar.
- Period: publications from 2015–2025.
- Search terms combined AI/SME/developing-economy/adoption/enablers/barriers/value-creation concepts.
- Iterative refinements and snowballing (forward/backward citation searches).
- Screening/selection:
- Initial results: 57 records → duplicates removed → 27 articles screened by title/abstract → full-text review using inclusion/exclusion criteria.
- Inclusion: empirical/theoretical/policy works focused on AI adoption by SMEs in developing economies (and comparable contexts), with practical relevance to implementation, enablers/barriers, or outcomes.
- Exclusion: studies solely about large firms, developed-economy–only studies without relevance, overly theoretical AGI-focused work, non-empirical opinion pieces, and pre-2015 material.
- Analytical framing: thematic synthesis using RBV, DC, TOE, and DOI to categorize internal resources, organizational adaptability, external environment, and adoption stages.
- Policy analysis: qualitative content analysis of Botswana national documents (Vision 2036, SmartBots, BDIH) to align policy intents with observed enablers/constraints.
Implications for AI Economics
- Microeconomic productivity and structural change:
- AI adoption in SMEs can raise firm-level productivity and potentially aggregate productivity, but net labor effects depend on skill-upgrade dynamics and task reallocation.
- Economists should quantify heterogeneous returns across sectors, firm sizes, and skill compositions.
- Market structure, competition, and rents:
- AI can increase differentiation and market power for adopters; policies should consider competition effects and barriers-to-entry created by uneven adoption.
- Finance & adoption dynamics:
- Limited access to finance is a key bottleneck; research should evaluate subsidy, credit, and blended-finance schemes’ cost-effectiveness in accelerating adoption.
- Data governance as economic infrastructure:
- Data rules affect spillovers, market creation, and the scope for platform-mediated SME growth; treating data governance as a public-good/infrastructure problem has distributional and innovation implications.
- Measurement and empirical strategy suggestions:
- Need for more firm-level microdata (panel surveys, administrative tax/utility data) tracking adoption, investment, revenue, employment, and productivity.
- Causal inference methods recommended: RCTs for training/finance interventions; difference-in-differences and synthetic controls for policy rollouts (e.g., SmartBots initiatives); instrumental variables for adoption endogeneity; structural models for diffusion and long-run general equilibrium effects.
- Study distributional impacts (winners vs losers), complementarities with human capital, and dynamic capabilities as mediating variables.
- Policy design insights for economists:
- Design targeted, evidence-based instruments: modular subsidies for capability building, matched finance tied to measurable adoption milestones, and public investment in shared digital infrastructure.
- Evaluate phased-adoption incentives and ecosystem subsidies (innovation hubs, co-working AI resources) to lower adoption fixed costs and spread network externalities.
- Open research questions:
- Magnitude and persistence of productivity gains from incremental vs. leapfrog AI adoption.
- Interaction between AI adoption and labor-skill upgrading in SMEs.
- Optimal public–private mixes for financing SME AI adoption in low-resource settings.
- How data governance regimes alter innovation incentives and cross-border SME market access.
If you want, I can convert these implications into a short research agenda with concrete empirical designs (e.g., an RCT outline, a diff‑in‑diff identification strategy using national program rollouts, or suggested survey modules for SME AI adoption).
Assessment
Paper Typereview_meta
Evidence Strengthlow — The paper is a structured narrative literature review synthesizing case studies, surveys, conceptual frameworks, and policy reports rather than new causal or quasi-experimental estimates; empirical evidence cited is heterogeneous and largely descriptive, so causal claims about long-run productivity, employment, and distributional impacts remain thin.
Methods Rigormedium — Uses established theoretical frameworks (RBV, DC, TOE, DOI) and a structured approach to synthesize cross-cutting enablers and barriers, but it is not a systematic review or meta-analysis, relies on varied study designs of uneven quality, and does not provide formal bias assessments or quantitative synthesis.
SampleA literature-based synthesis drawing on empirical studies (case studies, firm and sector surveys), conceptual frameworks, and policy reports with Botswana used as the primary illustrative context and broader relevance inferred for low- and middle-income countries; no original microdata or causal estimation is presented.
Themesadoption productivity skills_training innovation governance
GeneralizabilityFindings anchored in Botswana and selective LMIC examples may not generalize to countries with different institutions, infrastructure, or market structures, Heterogeneity across SME sectors, firm sizes, and firm informality limits cross-sector generalizability, Evidence relies on case studies and surveys that may suffer from selection bias and limited external validity, Rapid evolution of AI technologies and services reduces temporal generalizability; recommendations may require frequent updating, Limited empirical causal evidence on long-run welfare, employment, and distributional outcomes constrains policy transferability
Claims (28)
| Claim | Direction | Confidence | Outcome | Details |
|---|---|---|---|---|
| AI adoption by SMEs in developing economies (illustrated using Botswana) can materially enhance operational efficiency, customer personalization, innovation capacity, and competitive advantage, supporting sustainable economic diversification — but meaningful uptake is constrained by skills, infrastructure, finance, and fragmented data governance. Organizational Efficiency | mixed | medium | operational efficiency; customer personalization; innovation capacity; competitive advantage; AI adoption/uptake |
0.07
|
| Theoretical framing integrates Resource-Based View (RBV), Dynamic Capabilities (DC), Technology–Organization–Environment (TOE), and Diffusion of Innovation (DOI) to explain how firm resources, learning capacity, organizational and environmental factors shape AI adoption. Research Productivity | null_result | high | explanatory scope for AI adoption drivers (theoretical coherence rather than an empirical outcome) |
0.12
|
| AI can generate productivity gains for SMEs through automation and process optimization. Organizational Efficiency | positive | medium | productivity (e.g., output per worker, process cycle times, operational efficiency) |
0.07
|
| AI-enabled customer segmentation and personalization can increase sales and customer retention for SMEs. Firm Revenue | positive | medium | sales revenue; customer retention rates; conversion metrics |
0.07
|
| AI adoption supports new product/service innovation and faster time-to-market for SMEs. Innovation Output | positive | medium | number of new products/services; time-to-market (development cycle duration) |
0.07
|
| AI can improve market access for SMEs (e.g., via digital platforms and AI-enabled credit scoring) and enable potential value-chain upgrading. Market Structure | positive | medium | market access indicators (platform participation, sales channels); access to finance (credit approvals); position in value chain |
0.07
|
| Supportive government policy and adaptive regulation are important enablers of AI adoption among SMEs. Adoption Rate | positive | medium | AI adoption rate; regulatory environment quality |
0.07
|
| Local innovation ecosystems (universities, incubators, private-sector partnerships) support SME uptake of AI. Adoption Rate | positive | medium | formation of partnerships; technology transfer occurrences; AI adoption among SMEs |
0.07
|
| Access to finance, including scalable and blended financing models, is a key enabler for SME AI adoption. Adoption Rate | positive | medium | availability of tailored financing; uptake of AI investments by SMEs |
0.07
|
| External market pressures and customer demand often drive AI adoption decisions in SMEs. Adoption Rate | positive | medium | reported adoption triggers; AI adoption frequency linked to customer/market signals |
0.07
|
| Skills shortages (AI literacy, data science, digital management) are a primary constraint on SME AI adoption in developing economies. Skill Acquisition | negative | high | availability of AI-relevant skills; reported skills constraints limiting adoption |
0.12
|
| Infrastructure deficits (unreliable power, inadequate broadband, limited local compute) materially constrain AI uptake by SMEs. Adoption Rate | negative | high | infrastructure adequacy metrics (power reliability, broadband access); AI adoption incidence |
0.12
|
| High upfront costs and lack of tailored financing instruments are significant financial constraints on SME AI adoption. Adoption Rate | negative | high | upfront investment costs; access to tailored finance; adoption rates |
0.12
|
| Fragmented or weak data governance (privacy rules, standards, interoperability, and trust) reduces SMEs’ ability to participate in data-driven markets and adopt AI. Governance And Regulation | negative | medium | data governance quality; SME participation in data markets; trust/interoperability metrics |
0.07
|
| A phased adoption approach (assess needs → pilot low-risk use cases → scale modularly) is recommended to reduce risk and improve outcomes for SME AI projects. Adoption Rate | positive | medium | success rate of AI pilots; scalability of deployments; mitigation of adoption risks |
0.07
|
| Incremental investment in human capital and development of dynamic capabilities (learning, adaptation) increases SMEs’ absorptive capacity and the likelihood of successful AI adoption. Skill Acquisition | positive | medium | absorptive capacity metrics; successful AI adoption; firm performance post-adoption |
0.07
|
| Strengthening ecosystem linkages among academia, tech providers, financiers, and regulators enhances the prospects for inclusive, scalable AI adoption by SMEs. Adoption Rate | positive | medium | ecosystem connectivity; number of collaborative projects; SME AI uptake |
0.07
|
| Implementing scalable financing and procurement models (pay-as-you-go, leasing, blended finance) can overcome upfront cost barriers for SMEs adopting AI. Adoption Rate | positive | medium | use of alternative financing models; reduction in financing barriers; AI adoption rates |
0.07
|
| Developing pragmatic, locally appropriate data governance arrangements (standards, privacy safeguards, data trusts) is necessary to build trust and enable SME participation in data-driven markets. Governance And Regulation | positive | medium | trust in data sharing; interoperability; SME engagement in data ecosystems |
0.07
|
| Heterogeneity in SME types and sectors limits the generalizability of findings about AI adoption and impacts. Other | null_result | high | generalizability of reviewed findings across SMEs and sectors |
0.12
|
| Empirical causal evidence on long-run welfare, distributional outcomes, and labor effects of AI in LMIC SMEs remains thin. Research Productivity | null_result | high | availability of causal evidence on welfare, distributional effects, and labor outcomes |
0.12
|
| If AI diffusion is broad and SMEs possess absorptive capacity, AI can contribute to firm-level productivity improvements and sectoral diversification, potentially supporting aggregate growth; without capacity building, gains may concentrate among better-resourced firms. Firm Productivity | mixed | medium | firm productivity; sectoral diversification; distribution of gains across firm types |
0.07
|
| AI adoption shifts demand toward higher-skill tasks and complementary human capital, creating short-term displacement risks but opportunities for upskilling and higher-value employment if policies and training align. Employment | mixed | medium | employment composition changes; skill demand; displacement vs. upskilling outcomes |
0.07
|
| Platform effects and supplier ecosystems associated with AI may create winner-takes-most market dynamics, so policy should monitor market concentration and enable competitive access to core AI services. Market Structure | negative | medium | market concentration metrics; access to core AI services by SMEs |
0.07
|
| Innovative financing (blended finance, pay-per-use, outcome-linked financing) is critical to overcome upfront cost barriers and enable scalable, risk-sharing investments in AI for SMEs. Adoption Rate | positive | medium | uptake of innovative financing instruments; AI investment levels by SMEs |
0.07
|
| Robust, locally appropriate data governance (privacy, interoperability, standards) is a public good that underpins trust and data-driven markets; weak governance raises risks of exclusion and foreign dependency. Governance And Regulation | negative | medium | data governance robustness; SME inclusion in data-driven markets; foreign dependency indicators |
0.07
|
| Policy design should be adaptive and sector-sensitive, balancing innovation with safeguards while targeting skills, infrastructure, and inclusive finance to maximize social returns from SME AI adoption. Governance And Regulation | positive | medium | effectiveness of policy interventions; inclusive AI adoption metrics |
0.07
|
| Research priorities include causal studies on AI’s impacts on SME productivity, employment and inequality in LMICs; cost–benefit analyses of financing and policy interventions; evaluation of data governance models; and development of metrics/monitoring systems for inclusive adoption. Research Productivity | null_result | high | existence and quality of targeted causal and evaluative research on AI in LMIC SMEs |
0.12
|
Entities
Artificial Intelligence (ai_tool)
Small and Medium-sized Enterprises (population)
Botswana (population)
Structured narrative literature review (method)
Operational efficiency (outcome)
Productivity (outcome)
Customer personalization (outcome)
Innovation capacity (outcome)
Skills shortages (AI literacy, data science, digital management) (outcome)
Infrastructure deficits (power, broadband, compute access) (outcome)
Financial constraints (upfront costs, lack of tailored finance) (outcome)
Fragmented or weak data governance (privacy, standards, interoperability, trust) (outcome)
Resource-Based View (RBV) (method)
Dynamic Capabilities (DC) (method)
Technology-Organization-Environment (TOE) (method)
Diffusion of Innovation (DOI) (method)
Competitive advantage (outcome)
Market access (outcome)
Labor effects / Employment outcomes (outcome)
Low- and middle-income countries (LMICs) (population)
Developing economies (population)
Government / Public policy (institution)
Phased adoption / Strategic piloting and modular scaling (method)
Ecosystem collaboration (academia, tech providers, financiers, regulators) (method)
Adaptive regulation (method)
Blended finance (method)
Distributional outcomes (outcome)
Inequality (outcome)
Case studies (method)
Surveys (dataset)
Policy reports (dataset)
Universities (institution)
Incubators (institution)
Private sector partnerships (institution)
Regulators (institution)
Financiers / Finance providers (institution)
Digital platforms (institution)
Pay-as-you-go financing (method)
Outcome-linked financing (method)
Data trusts (method)
Federated learning (method)
Causal impact studies (method)
Cost–benefit analysis (method)
Metrics and monitoring systems for inclusive adoption (method)
Leasing (technology leasing models) (method)