Firms that adopt AI strategically report stronger performance, because AI appears to boost strategic agility; the evidence is based on a 220‑firm cross‑sectional survey and shows association rather than proven causation.

Main Finding

Strategic adoption of AI positively affects firm performance in an emerging-economy sample (Iraq), and a substantial part of this effect operates indirectly: strategic AI adoption strongly increases firms’ strategic agility, which in turn raises performance. Mediation is partial—AI has both a direct and an agility-mediated effect on performance.

Key Points

• Sample & context: N = 220 firms in Iraq across banking/finance (32.7%), telecommunications (18.6%), services (30.5%), and manufacturing (18.2%). Respondents were managerial staff; data collected by online survey (purposive + snowball sampling).
• Hypotheses tested and results:
  • H1 (SAI → Firm Performance): supported. β = 0.21, t = 2.45, p = 0.015.
  • H2 (SAI → Strategic Agility): supported. β = 0.52, t = 8.10, p < 0.001.
  • H3 (Strategic Agility → Firm Performance): supported. β = 0.43, t = 5.60, p < 0.001.
  • H4 (mediation): strategic agility partially mediates the SAI → performance link. Indirect effect β = 0.22, t = 4.30, p < 0.001.
• Explanatory power: R² = 0.27 for strategic agility; R² = 0.41 for firm performance (moderate to substantial).
• Measurement quality: strong reliability and validity—Cronbach’s α (SAI 0.87; SAG 0.85; FP 0.89), composite reliabilities (0.90–0.92), AVE (0.62–0.66); HTMT discriminant validity all < 0.85.
• Measures: reflective multi-item scales, 5-point Likert. Firm performance measured perceptually relative to competitors over the past 3 years.
• Analysis: PLS-SEM (SmartPLS), bootstrapping with 5,000 subsamples.

Data & Methods

• Design: Cross-sectional quantitative survey.
• Population & unit: Firms operating in Iraq; unit of analysis = firm; respondents were managers, heads of department, directors, supervisors.
• Sampling: Purposive and snowball sampling; 250 questionnaires distributed → 220 valid responses retained.
• Constructs:
  • Strategic AI Adoption (extent AI is embedded into strategy and decision processes; 5 items).
  • Strategic Agility (sensing, reconfiguration, rapid strategic adjustment; 5 items).
  • Firm Performance (6 items: profitability, growth, market share, customer outcomes; perceptual).
• Measurement model checks: item loadings ≳0.77; Cronbach’s α, Composite Reliability, AVE reported and acceptable; HTMT < 0.85.
• Structural model checks: VIF (collinearity diagnostics reported), path coefficients, significance via bootstrap (5,000), R² reported, mediation tested using indirect effects.
• Limitations (methodological): cross-sectional design limits causal inference; non-probability sampling limits strict generalizability; performance measured perceptually (no objective financials).

Implications for AI Economics

• Mechanisms of value creation: The paper provides empirical support that AI increases firm performance both directly (likely through efficiency and decision-quality gains) and indirectly by strengthening firms’ strategic agility (improving sensing, reconfiguration and strategic responsiveness). For economists, this underscores the importance of capability complementarities: AI is not a freestanding productivity input but yields larger returns when paired with organizational capabilities.
• Policy and diffusion in emerging economies: Results suggest policymakers in emerging markets should promote not only AI adoption but also complementary investments—management training, organizational redesign, and institutions that reduce uncertainty—so firms can realize strategic benefits. Support measures (subsidies, advisory services) that encourage strategic alignment of AI may raise aggregate productivity gains from digitalization.
• Firm heterogeneity and market structure: The moderate R² and partial mediation imply heterogeneity in returns—firm size, sector, managerial capability, and resource endowments likely modulate the AI → agility → performance chain. Economic analyses of AI diffusion should model heterogeneous adoption payoffs and complementarities across firm types.
• Measurement and macro inference: Reliance on perceptual performance highlights an empirical gap: macro- and industry-level estimates of AI’s productivity effects should combine firm-reported measures with objective outcomes (financials, output, employment) to quantify economy-wide impacts and labor-market consequences.
• Research priorities for AI economics:
  • Causal identification: longitudinal/panel designs or natural experiments to identify causal impacts of strategic AI adoption on productivity.
  • Complementarity estimation: formal models and microdata estimating complementarities between AI and organizational capital (skills, processes).
  • Distributional effects: examine how AI-driven agility affects firm growth, market concentration, and labor reallocation in emerging markets.
  • Sectoral/scale thresholds: identify minimal capability thresholds (firm size, human capital) required for AI to generate strategic returns.
• Practical takeaway for economic modeling: When integrating AI into production-function or growth models, include an endogenous organizational-capability channel (strategic agility or similar) to capture amplification or attenuation of AI’s productivity effects.