AI investments pay off mainly by speeding decision-making: firms' technical and managerial AI capabilities raise performance chiefly through greater decision-making agility, but the payoff depends on how dynamic and complex the external environment is; configurational analysis shows multiple routes to high performance depending on firm and environmental conditions.

AI for decision-making: exploring the linkage from AI capability to firm performance

Boram Kwon, Junyeong Lee, Hyungjin Lukas Kim · Fetched May 24, 2026 · Aslib Journal of Information Management

semantic_scholar correlational low evidence 7/10 relevance DOI Source

Using a survey of 251 firms, the paper finds that AI technical infrastructure and management improve firm performance indirectly by enhancing decision-making agility, and that environmental dynamism and complexity alter the strength of this linkage, while fsQCA reveals multiple capability configurations associated with high performance.

Artificial intelligence (AI) stands as a pivotal, revolutionary force in technologically reshaping industries. Despite extensive research on the potential of AI, the specific mechanisms through which AI capabilities lead to competitive advantages still need delineated. To address this notable gap in literature, we investigate how agility in decision-making, as a focal dynamic capability, is a critical conduit linking AI capabilities for improving organizational outcomes and examine the interplay between a firm's decision-making agility and its internal and external environment, thereby offering a comprehensive understanding of the dynamics of today's complex business ecosystem. The mixed-method approach, combining partial least squares–structural equation modeling (PLS-SEM) and fuzzy-set qualitative comparative analysis (fsQCA), was used for analyzing the survey data of 251 firms. The results of PLS-SEM show that two sub-dimensions of AI capability, technical infrastructure and management, affect performance outcomes through decision-making agility. Moreover, environmental dynamism and complexity differently moderate the relationship between decision-making agility and firm performance. Additionally, the results of fsQCA demonstrate how the combination and roles of strategic resources (e.g. AI capabilities and decision-making agility) shift in response to varying organizational and environmental conditions. By elucidating these dynamics via a mixed-method approach, our findings not only offer a deeper understanding of the strategic value of AI within the organizational dynamic capability perspective, but also provide practical insights for AI implementation that prioritizes management capability and adaptability to external environments.

Summary

Main Finding

Decision-making agility is the key channel through which AI capabilities (especially technical infrastructure and AI management) translate into improved firm performance, and the strength of this channel depends on environmental dynamism and complexity — different environmental conditions change which combinations of AI capabilities and agility lead to superior outcomes.

Key Points

AI capability is multi-dimensional; two critical sub-dimensions identified are:
- Technical infrastructure (tools, data architectures, computational resources).
- AI management (governance, processes, managerial competence to deploy/use AI).
Decision-making agility (speed, flexibility, and quality of decisions) mediates the effect of AI capabilities on firm performance.
Environmental moderators:
- Environmental dynamism (rapid, unpredictable change) and environmental complexity (many interacting variables) moderate how strongly agility translates into performance, but they do so differently (i.e., their effects are not identical).
Methodological complementarity:
- PLS-SEM establishes net effects, mediation, and moderation relationships.
- fsQCA uncovers multiple sufficient configurations (combinatorial paths) showing how different mixes of AI capability, agility, and contextual factors can lead to high performance.
Practical emphasis: managerial capabilities and adaptability to context are as important as technical investments for realizing the strategic value of AI.

Data & Methods

Sample: Survey data from 251 firms.
Mixed-method approach:
- Partial least squares–structural equation modeling (PLS-SEM) to test hypothesized relationships (mediation by decision-making agility; moderation by environmental dynamism and complexity).
- Fuzzy-set qualitative comparative analysis (fsQCA) to identify alternative configurations of resources and conditions that lead to high performance, highlighting equifinality and context-dependent roles of capabilities.
Analytic focus: Interaction of internal capabilities (AI infra + management, decision agility) with external environmental characteristics.

Implications for AI Economics

Mechanism clarity: Shows how AI investments convert into economic returns mainly via improved decision-making processes, refining models of AI-driven productivity beyond treating AI as a black-box capital input.
Complementarities matter: Returns to AI depend on complementary managerial capabilities and organizational agility — economic models of AI diffusion should include complementarities and complementarizing investments.
Heterogeneous gains across environments: Policy and firm-level cost–benefit analyses should account for environmental dynamism and complexity; same AI investment can yield different marginal returns depending on context.
Strategic investment guidance: Firms should balance spending between technical infrastructure and managerial capability building (training, governance, processes) to maximize ROI from AI.
Firm heterogeneity and equifinality: Multiple paths to superior performance exist; economists modeling industry-level effects of AI should allow for heterogeneous firm strategies and non-linear, configuration-dependent outcomes.
Future research directions for AI economics: quantify how much managerial/organizational capital is required per unit of AI capital for given environments; estimate general equilibrium effects when widespread adoption requires complementary investments.

Assessment

Paper Typecorrelational Evidence Strengthlow — Findings are based on a single cross-sectional survey (n=251) with self-reported measures and correlational analyses; while mediation and moderation are estimated, the design cannot rule out reverse causation, omitted variable bias, or common-method bias, so causal interpretation is weak. Methods Rigormedium — The authors use a mixed-methods approach (PLS-SEM for path analysis and fsQCA for configurational insights), which is appropriate for exploring complex relationships and heterogeneity; however, rigor is limited by moderate sample size, unclear sampling frame/representativeness, reliance on cross-sectional self-report data, and no robustness checks (e.g., longitudinal analysis, instrumental variables, or objective performance measures) reported. SampleSurvey data from 251 firms (details on country, industries, firm-size distribution, respondent roles, and sampling procedure not specified in the summary); variables include AI capability sub-dimensions (technical infrastructure, management), decision-making agility, firm performance (self-reported), and moderators for environmental dynamism and complexity. Themesorg_design innovation IdentificationObservational cross-sectional survey analysis using PLS-SEM to estimate associations including mediation (AI capabilities -> decision-making agility -> performance) and moderation (environmental dynamism/complexity on agility->performance), supplemented by fsQCA to identify configurational patterns; no exogenous variation, random assignment, panel/time-series, or instrumental variables to support causal claims. GeneralizabilityModerate-to-small sample size (n=251) limits statistical power and external validity, Unknown geographic and industry coverage — potentially single-country or convenience sample, Findings rely on self-reported firm performance, raising measurement and common-method bias concerns, Cross-sectional design limits inference about temporal or causal ordering, fsQCA configurations may be sensitive to calibration choices and sample-specific idiosyncrasies

Claims (7)

Claim	Direction	Confidence	Outcome	Details
The mixed-method approach, combining partial least squares–structural equation modeling (PLS-SEM) and fuzzy-set qualitative comparative analysis (fsQCA), was used for analyzing the survey data of 251 firms. Other	null_result	high	research methodology / analytic approach	n=251 0.5
Two sub-dimensions of AI capability, technical infrastructure and management, affect performance outcomes through decision-making agility. Firm Productivity	positive	high	firm performance (through decision-making agility)	n=251 0.3
Decision-making agility is a critical conduit linking AI capabilities to improving organizational outcomes. Firm Productivity	positive	high	organizational/firm performance mediated by decision-making agility	n=251 0.3
Environmental dynamism and complexity differently moderate the relationship between decision-making agility and firm performance. Firm Productivity	mixed	high	moderation of decision-making agility effect on firm performance by environmental dynamism and complexity	n=251 0.3
The results of fsQCA demonstrate how the combination and roles of strategic resources (e.g. AI capabilities and decision-making agility) shift in response to varying organizational and environmental conditions. Innovation Output	mixed	high	configurations (combinations) of resources associated with firm performance under varying conditions	n=251 0.3
Findings provide practical insights for AI implementation that prioritize management capability and adaptability to external environments. Organizational Efficiency	positive	high	organizational effectiveness of AI implementation (management capability and adaptability as priorities)	n=251 0.05
AI capability is conceptualized/measured as having sub-dimensions including technical infrastructure and management. Other	null_result	high	construct dimensionality of AI capability	n=251 0.3