Buying AI tools is not enough: leading firms convert AI investment into measurable value through a constellation of eight organizational capabilities — driven by top management and multi‑year coordination — rather than by raw spending alone.

Main Finding

Companies’ ability to convert AI investment into tangible results depends less on raw spending on tools and more on a broader set of organizational capabilities. Using interviews with CIOs and an abductive Capability View analysis, the authors show that AI deployment is a multi-year, multi-competency process driven by top-management leadership, and they propose an AI Deployment Capability Framework (AI‑DCF) that identifies eight critical capabilities explaining variation in strategies, approaches, and outcomes across firms.

Key Points

• AI deployment remains at an early and uneven stage across leading publicly listed firms; interest and investment have outpaced realized benefits.
• Progress depends on a broader constellation of capabilities than most prior peer‑reviewed work has emphasized.
• The authors identify and define eight critical capabilities (aggregate, cross‑company differences explained by these capabilities), and package them into the AI‑DCF as a practical and researchable framework.
• AI deployment is characterized as a comprehensive, multi‑year effort requiring coordination across competencies rather than a one‑off technology purchase.
• Top management plays a vital role in guiding, prioritizing, and sustaining AI capability development.
• This study is presented as the first management‑level empirical identification and explanation of the essential capabilities needed to achieve measurable AI benefits.

Data & Methods

• Empirical base: semi‑structured interviews with chief information officers at leading international publicly listed companies (sample details such as number and sectors not provided in the summary).
• Analytical approach: abductive analysis (iterating between data and theory) using the Capability View as the theoretical lens to uncover patterns and explain differences in deployment outcomes.
• Output: the AI‑DCF — a strategy‑based framework that synthesizes the identified capabilities and situates AI deployment as a long‑horizon organizational transformation.

Implications for AI Economics

• Heterogeneous returns to AI: Observed variation in realized benefits likely reflects differences in firm‑level capability stocks, not just differences in investment levels or access to tools. Empirical work that treats AI investment as uniform risks biased inference.
• Complementarities matter: Organizational capabilities (management, processes, data governance, skills, etc.) are complements to AI capital. Productivity gains will depend on these complementarities, consistent with models of intangible capital and organizational capital.
• Measurement and modeling: Economists should measure capability dimensions (using frameworks like AI‑DCF) and include them as covariates or structural state variables in production‑function and adoption models to better estimate returns to AI.
• Policy and firm strategy: Policies or programs that subsidize AI tools alone may have limited impact; interventions that build managerial capacity, data infrastructure, workforce skills, and governance may be more effective. Firms should plan AI as a multi‑year capability accumulation process led from the top.
• Research agenda: Use the AI‑DCF to develop survey instruments and panel measures of firm capabilities; investigate how capability accumulation interacts with labor outcomes, reallocation, and firm productivity dynamics; design causal identification strategies (e.g., panel approaches, instruments, natural experiments) that account for capability heterogeneity and path dependence.