Across several Central and Eastern European economies, AI adoption rarely boosts output unless firms already have standardized processes, management systems and skilled workers; without these complements, AI gets filtered out in a 'productivity funnel'. Policy and investment should prioritize organizational readiness and skills alongside AI tools to realize productivity gains.

Main Finding

Widespread AI adoption has produced only modest productivity gains because technological potential is filtered out by organizational constraints. AI increases productivity meaningfully only in firms that possess adequate organizational readiness—standardized processes, management systems, and complementary human capital. In many Central and Eastern European (CEE) firms, particularly smaller firms and those in Serbia, Croatia, Czechia, and Romania, AI diffusion alone does not translate into value capture: firms get stuck in a “productivity funnel” unless they invest in organizational complements.

Key Points

• Productivity funnel: technological potential narrows through stages—(1) access and digital infrastructure, (2) organizational absorption, (3) human capital adaptation, and (4) ultimate value capture. Each stage is a filter; failure at any stage prevents productivity gains.
• Complementarity trap: firms with low organizational readiness adopt AI but cannot exploit it; they remain stuck and fail to convert diffusion into productivity improvements.
• Conditional effects: AI adoption has no robust direct productivity effect. Positive productivity impacts appear only when firms have pre-existing management practices, standardized processes, and relevant human capital.
• Heterogeneity: AI adoption rates and the size of conditional productivity effects vary by country and firm size. Small and less-organized firms show the weakest productivity payoffs.
• Interpretation of the AI productivity paradox: The paradox is driven more by organizational and absorptive-capacity constraints than by intrinsic technological limits of AI.

Data & Methods

• Data: Firm-level microdata from a subset of CEE economies—Serbia, Croatia, Czechia, and Romania—linked with indicators of AI diffusion (e.g., AI tool adoption, AI-related expenditures, or AI-intensity proxies).
• Measurement of organizational readiness: Indicators such as presence of standardized processes, formal management systems, quality assurance, training programs, and workforce skill composition.
• Empirical strategy:
  • Estimate productivity (e.g., TFP or labor productivity) regressions including AI adoption measures.
  • Interact AI adoption with organizational readiness indicators to test conditional effects.
  • Control for firm characteristics (size, age, sector), country fixed effects, and observable confounders; implement robustness checks (alternative readiness metrics, subsamples by size/sector).
• Identification: Relies on cross-sectional and panel variation in adoption and readiness; causal interpretation strengthened by robustness tests and by showing heterogeneous effects aligned with theoretical predictions (i.e., gains only where complements exist).

Implications for AI Economics

• Rethink the productivity paradox: Research and policy should emphasize complementary organizational capital and absorptive capacity as primary constraints to AI-driven productivity growth.
• Measurement: Empirical studies must measure organizational readiness and heterogeneity explicitly; average treatment effects of AI adoption can be misleading without conditioning on complements.
• Policy sequencing: Policies should prioritize investments in management practices, process standardization, workforce training, and change management—especially for SMEs—before or alongside subsidies for AI tools.
• Firm strategy: Firms should assess and build organizational complements (processes, governance, skills) to capture AI value; piecemeal AI procurement without internal adaptation is unlikely to raise productivity.
• Targeting and equity: Support programs in early-adopting economies should focus on smaller firms and lagging regions to avoid widening productivity gaps driven by organizational capacity differences.
• Future research directions: Longitudinal studies of organizational change post-AI adoption, industry- and task-level analyses of where complementarities matter most, and evaluation of policies that build organizational readiness.