Firms that embed AI and stay strategically agile report stronger sustainable performance, aided by improved dynamic capabilities and a data-driven culture; proactive corporate governance further amplifies gains from dynamic capabilities.

Main Finding

AI integration and strategic agility both positively and significantly increase sustainable firm performance in Indian IT/ITeS firms, but their value is substantially realized through organizational complements — dynamic capabilities and a data‑driven culture — and is further strengthened when corporate governance is proactive. The model explains 58.7% of performance variance (R² = 0.587).

Key Points

• Direct effects
  • AI Integration → Sustainable Firm Performance: β = 0.28, p < 0.001.
  • Strategic Agility → Sustainable Firm Performance: β = 0.35, p < 0.001 (largest direct effect).
• Mediation (indirect paths are significant but smaller)
  • Dynamic Capabilities mediate AI and Agility effects (AI→DC→Performance β ≈ 0.11; Agility→DC→Performance β ≈ 0.13; both p ≤ 0.001).
  • Data‑Driven Culture also mediates both relationships (AI→DDC→Performance β ≈ 0.08, p = 0.004; Agility→DDC→Performance β ≈ 0.11, p < 0.001).
• Moderation
  • Proactive Corporate Governance amplifies the positive effect of Dynamic Capabilities on performance (interaction β = 0.15, p ≤ 0.001).
  • It also strengthens the Data‑Driven Culture → performance path (interaction β = 0.09, p = 0.015).
• Measurement quality
  • Sample: 327 senior and mid‑level managers in Indian IT/ITeS.
  • Instruments: validated multi‑item reflective scales; Cronbach’s α and composite reliabilities all strong (α ≈ 0.885–0.934; AVE > 0.60).
  • Predictive relevance: Q² = 0.432.
• Limitations noted by authors: cross‑sectional design (limits causal claims), single sector and country (generalizability).

Data & Methods

• Design: Cross‑sectional survey (purposive sampling of managers); usable n = 327 (response rate 72.6% from 450).
• Context: Indian IT and IT‑enabled services firms.
• Constructs measured (7‑point Likert); sources: prior validated scales for AI capability, strategic agility, dynamic capabilities, data‑driven culture, proactive corporate governance, and sustainable firm performance (financial/market/operational over 3 years).
• Analysis: PLS‑SEM (SmartPLS 4.0); two‑step approach (measurement then structural model); bootstrapping with 5,000 subsamples; collinearity checks (VIF < 3), Fornell‑Larcker for discriminant validity.
• Key model statistics: R²(SFP) = 0.587; Q² = 0.432; mediation and moderation tested simultaneously.

Implications for AI Economics

• Complementarities matter: The paper provides empirical support for the classic economics insight that technology (AI) is productive only with complementary investments. Estimates suggest sizable direct returns to AI but meaningful additional value through organizational complements (dynamic capabilities, culture). Empirical work on AI’s economic impact should control for these complements to avoid biased estimates.
• Bundling investments: Firms (and policymakers incentivizing AI adoption) should treat AI as part of a bundle — data infrastructure + organizational processes + governance — not as a standalone capital investment. Cost‑benefit analyses and ROI models must include spending on capability building and culture change.
• Role of governance: Proactive corporate governance (board-level digital literacy, strategic oversight, ethics) amplifies returns. From a regulatory or firm‑governance perspective, interventions that raise governance quality may increase the social returns to AI adoption by reducing misuse and misalignment.
• Measurement & identification guidance: The study highlights measurable mediators (dynamic capabilities, data‑driven culture) that can be included as controls or mechanisms in productivity and labor‑market impact studies of AI. Researchers seeking causal estimates should prioritize longitudinal designs and instrumenting for complementarities to separate adoption effects from capability accumulation.
• Distributional and structural considerations: If returns to AI are conditional on organizational capabilities and governance, diffusion may be uneven — widening productivity gaps between firms that can build these complements and those that cannot. This heterogeneity matters for aggregate productivity, employment displacement/creation, and competition policy.
• Future empirical priorities for AI economics
  • Estimate causal returns using panel/experimental designs, accounting for time to build capabilities.
  • Quantify the relative cost of achieving the complements (training, process redesign, governance upgrades) versus direct AI investment.
  • Model equilibrium effects: how firm‑level complementarities affect industry dynamics, wages, and inequality during AI diffusion.

If you want, I can (a) extract the exact survey items / scales used in each construct, (b) produce a short figure summarizing the reported path coefficients, or (c) outline how to operationalize these constructs in a productivity study. Which would be most helpful?