← Papers

A systematic review of 72 studies shows SMEs adopt AI mainly in manufacturing and in Asia, led by technological, organizational and leadership factors; contrary to expectations, environmental uncertainty often amplifies AI’s benefits—boosting performance, innovation and resilience—though evidence rests chiefly on cross‑sectional studies.

Deciding Under Uncertainty: A Systematic Review and Integrated Theories–Contexts–Methods and Antecedents– Decisions–Outcomes Framework of Artificial Intelligence-Driven Decision-Making in Small and Medium-Sized Enterprises

Omar Picone Chiodo, Mazumder Sita, Noptanit Chotisarn, Thadathibesra Phuthong · September 04, 2026 · International Review of Management and Marketing

openalex review_meta n/a evidence 7/10 relevance Full text usable extracted full text DOI Source PDF

Structured author observations

Linked only from stored provider relations; the raw author line above is never matched by name.

OpenAlex

Latest observation: September 7, 2026

Omar Picone Chiodo provider ID
Mazumder Sita provider ID
Noptanit Chotisarn provider ID
Thadathibesra Phuthong provider ID

A systematic review of 72 Scopus‑indexed studies (2018–2025) finds that AI adoption in SMEs is driven by technological, organizational, environmental, and leadership antecedents, that adoption concentrates in manufacturing and Asia, and that environmental uncertainty tends to amplify AI’s benefits for performance, innovation, sustainability, and resilience.

Citation observations

Cumulative provider counts captured on specific dates; providers are never combined.

0 cumulative citations

OpenAlex · Observed September 7, 2026

View corpus context

0 cumulative citations

Semantic Scholar · Observed September 7, 2026

View corpus context

significant knowledge gaps remain regarding how small and medium-sized enterprises (SMEs) leverage AI under volatile business conditions. This study systematically reviews AI-driven decision-making in SMEs operating under environmental uncertainty. Following the Scientific Procedures and Rationales for Systematic Literature Reviews (SPAR-4-SLR) protocol, we analyze 72 peer-reviewed articles (2018–2025) and develop an integrated theories–contexts–methods and antecedents–decisions–outcomes (TCM–ADO) framework. The findings show that manufacturing is the dominant research context and that Asia, especially China, accounts for 55% of studies, with quantitative cross-sectional surveys the prevailing methodology. AI adoption is shaped by technological, organizational, environmental, and leadership antecedents that influence strategic, operational, financial, and human–AI decision processes; the resulting outcomes span business performance, innovation, sustainability, and resilience. Notably, environmental uncertainty amplifies rather than diminishes AI benefits, positioning AI as an adaptive mechanism during turbulence rather than a barrier. The review contributes an integrated framework that connects how the phenomenon is studied with what is substantively known about it, and it offers a structured agenda for future work. For practitioners, the findings underscore the value of treating AI strategically, building complementary capabilities, and maintaining flexible organizational structures.

Summary

Main Finding

AI adoption in SMEs under environmental uncertainty is driven by technological, organizational, environmental, and leadership antecedents and primarily improves firms’ strategic, operational, financial, and human–AI decision processes — producing gains in performance, innovation, sustainability, and resilience. Crucially, environmental uncertainty tends to amplify AI’s benefits (AI functions as an adaptive mechanism in turbulence), rather than primarily acting as a barrier. The authors synthesize these results into an integrated TCM–ADO (theories–contexts–methods and antecedents–decisions–outcomes) framework and a structured research agenda.

Key Points

Scope and corpus
- Systematic review of 72 peer‑reviewed empirical articles (2018–2025) from Scopus.
- Focus: AI-driven decision-making in small and medium-sized enterprises (SMEs) operating under environmental uncertainty.
Theoretical landscape
- Dominant theories: Resource-Based View (RBV; 17 articles) and Dynamic Capabilities Theory (DCT; 15).
- Other common frames: TOE (technology–organization–environment; 10), Information Processing Theory, TAM, institutional and behavioral theories.
Research contexts & geographic distribution
- Top industry contexts: Manufacturing (11), Technology adoption studies (10), Supply chain (8), Digital transformation (7).
- Geography concentrated in Asia (55% of studies); China accounts for ~23% (14 articles). Cross‑national studies ~16.7%; Europe ~15%.
Methods
- Methodological dominance of quantitative cross-sectional surveys; limited longitudinal, experimental, or mixed-method designs.
- Studies spread across 46 journals; most frequent outlets: Sustainability (10), Technological Forecasting and Social Change (8), Technology in Society (5).
Antecedents, decisions, and outcomes (ADO)
- Antecedents: technological (IT/data infrastructure, AI tools), organizational (readiness, routines, dynamic capabilities), environmental (market volatility, regulatory change), leadership and managerial support.
- Decisions influenced: strategic (market positioning), operational (production, supply chain), financial (investment/price decisions), human–AI (task allocation, human oversight).
- Outcomes: improved business performance, innovation output, sustainability metrics, and organizational resilience.
Key empirical insight
- Environmental uncertainty often amplifies AI value: AI helps SMEs process data faster, reduce bias, and adapt in volatile settings; it is more often an enabler than a constraint during turbulence.
Limitations identified by authors
- Single-database (Scopus) search may omit some studies.
- Exclusion of conceptual/review pieces and conference proceedings.
- Geographic concentration and method mix limit generalizability and causal inference.

Data & Methods (of the review)

Review protocol: SPAR-4-SLR (assembling, arranging, assessing).
Search parameters: Scopus, 2018–2025, English-language journal articles; Boolean string combining AI, decision-making, SME, performance, and uncertainty terms.
Screening: 1,501 initial hits → title/abstract → 184 → full-text → 78 → final coded corpus of 72 empirical articles (cross-referenced and hand-checked).
Coding framework: combined TCM (theories–contexts–methods) with ADO (antecedents–decisions–outcomes); dual coding with consensus reconciliation.
Analysis: descriptive mapping (publication trends, outlets, geography, theories, contexts) and inductive thematic coding (antecedents, decisions, outcomes).

Implications for AI Economics

Productivity and firm heterogeneity
- AI adoption enhances SME decision quality and operational resilience, suggesting potential productivity gains for a large segment of the economy (SMEs comprise ~90–95% of firms globally). Heterogeneous access to AI (resources, capabilities, leadership) will likely widen productivity dispersion across firms.
Returns to AI under uncertainty
- Environmental uncertainty amplifies AI benefits, implying higher marginal returns to AI investment in volatile markets. Economic models of technology adoption should incorporate uncertainty as a positive moderator of adoption payoffs.
Market structure and competition
- If AI disproportionately raises resilience and strategic agility for adopters, markets could see accelerated competitive sorting: better-capitalized or better-managed SMEs may consolidate advantage, potentially altering entry/exit dynamics in turbulent industries.
Labor and task allocation
- AI-altered human–AI decision processes point to changes in task allocation and skill demand within SMEs. Microeconomic models should account for reallocation effects and complementarities between human skills and AI.
Policy and public investment
- Given resource constraints of many SMEs, policy interventions (subsidies for data/IT infrastructure, training, shared AI services) could be welfare-improving by reducing adoption frictions and asymmetric access to adaptive technologies.
Empirical and measurement needs
- The field’s methodological skew toward cross-sectional surveys limits causal inference. AI economics would benefit from longitudinal firm-level microdata linking AI adoption/use to output, productivity, employment, and investment dynamics — especially across different levels of market uncertainty.
External validity and global context
- The geographic concentration in Asia (notably China) suggests context-specific drivers and outcomes; comparative and cross-country work is needed before generalizing implications for other institutional and market environments.
Research recommendations relevant to AI economists
- Develop structural or reduced-form econometric analyses that estimate causal effects of AI adoption under different uncertainty regimes.
- Build models of adoption investment under uncertainty that incorporate dynamic capabilities and learning-by-doing for SMEs.
- Evaluate targeted policies (grants, shared AI platforms, training) using experimental or quasi-experimental designs to measure welfare and distributional impacts.
- Link firm-level AI usage metrics (not only adoption) to productivity, risk exposure, and survival in turbulence.

Summary judgment: The review synthesizes a nascent but fast-growing literature showing that AI can be a meaningful adaptive technology for SMEs facing environmental uncertainty. For AI economics, this underscores the importance of modeling uncertainty-dependent returns to AI, measuring heterogeneous impacts across firms, and crafting policies to equalize access and capture broader productivity gains.

Assessment

Paper Typereview_meta Evidence Strengthn/a — This is a systematic literature review synthesizing 72 empirical studies rather than presenting new causal estimation; the underlying body of evidence is dominated by cross‑sectional observational studies with limited causal identification, so the paper does not itself establish causal claims. Methods Rigormedium — The review follows an established SPAR‑4‑SLR protocol, uses a structured TCM–ADO coding framework, reports screening steps, and used dual coding with reconciliation; however it relied on a single database (Scopus), excluded non‑English and non‑journal items (e.g., conference papers, gray literature), did not report interrater reliability statistics, and the included primary studies are methodologically heterogeneous (many cross‑sectional surveys). SampleSystematic review of 72 peer‑reviewed empirical articles indexed in Scopus from 2018–2025 addressing AI‑driven decision‑making in SMEs under environmental uncertainty; studies span 46 journals across management, information systems, engineering and related fields; primary studies are predominantly quantitative cross‑sectional surveys with some qualitative work; manufacturing contexts and Asian countries (especially China) are over‑represented. Themesadoption productivity human_ai_collab org_design GeneralizabilitySearch limited to a single database (Scopus) may have omitted relevant studies indexed elsewhere (Web of Science, Google Scholar) or in conference proceedings., Restricted to English‑language, peer‑reviewed journal articles — excludes gray literature and potentially relevant conceptual or practitioner work., Geographic concentration in Asia (notably China) limits transferability to other institutional/economic contexts., Sectoral skew toward manufacturing, supply chain, and technology adoption contexts reduces applicability to services or other industries., Time window (2018–2025) focuses on recent work but omits earlier foundational studies., Primary studies’ heavy reliance on cross‑sectional designs constrains causal inference and longitudinal understanding.

Claims (9)

Claim	Direction	Outcome	Confidence & Evidence	Details
Across the reviewed literature, AI adoption in SMEs is shaped by technological, organizational, environmental, and leadership antecedents. Adoption Rate	mixed	Factors associated with SME adoption of AI-driven decision-making	Reading fidelity high Study strength medium	n=72 0.24
The reviewed studies link AI-driven decision-making in SMEs to strategic, operational, financial, and human–AI decision processes. Decision Quality	positive	Use of AI in organizational decision processes	Reading fidelity high Study strength medium	n=72 0.24
The review finds that the outcomes associated with AI-driven decision-making in SMEs include business performance, innovation, sustainability, and resilience. Firm Productivity	positive	Business performance, innovation, sustainability, and organizational resilience	Reading fidelity high Study strength medium	n=72 0.24
Environmental uncertainty amplifies rather than diminishes the benefits associated with AI-driven decision-making in SMEs. Organizational Efficiency	positive	Benefits of AI-driven decision-making under environmental uncertainty	Reading fidelity high Study strength low	n=72 0.12
AI is positioned in the reviewed literature as an adaptive mechanism that can help SMEs respond during periods of environmental turbulence. Organizational Efficiency	positive	SME adaptability and resilience during environmental turbulence	Reading fidelity high Study strength low	n=72 0.12
Manufacturing is the dominant empirical context in the reviewed research, where AI is reported to strengthen resilience and reduce production uncertainty. Organizational Efficiency	positive	Manufacturing resilience and production uncertainty	Reading fidelity high Study strength medium	n=11 11 articles 0.24
SMEs face financial constraints, knowledge gaps, and limited organizational readiness that can impede adoption of advanced AI technologies. Adoption Rate	negative	SME capability to adopt advanced AI technologies	Reading fidelity high Study strength low	not reported 0.12
AI can improve information processing, reduce cognitive bias, and increase decision speed and accuracy under conditions of environmental uncertainty. Decision Quality	positive	Decision speed, decision accuracy, information processing, and cognitive bias	Reading fidelity high Study strength low	not reported 0.12
Dynamic capabilities theory is used in the reviewed literature to explain how AI adoption can strengthen operational performance by enabling firms to integrate, build, and reconfigure capabilities in response to environmental change. Firm Productivity	positive	Operational performance and capability reconfiguration	Reading fidelity high Study strength low	n=15 15 articles 0.12