Building AI Capability from Digital Orientation: The Roles of Organizational Learning, Organizational Forgetting, and Trust in AI

The burgeoning importance of artificial intelligence (AI) in the business landscape is undeniable, yet a significant gap persists in empirical research detailing how companies translate their digital orientation into tangible AI capabilities. This disconnect is particularly stark as many firms invest heavily in digital and AI initiatives, only to falter in establishing a coherent and effective capability base. A recent study, leveraging data collected in 2025 from 306 Chinese firms, proposes a nuanced framework to illuminate this complex transformation. The research suggests that a firm’s digital orientation influences its AI capability through two key organizational processes: learning and forgetting, with the level of trust in AI systems acting as a crucial moderating factor.

The AI Adoption Paradox: Widespread Use, Limited Impact

The findings underscore a growing paradox in the AI adoption journey. While an estimated 88% of organizations globally now utilize AI in at least one business function, a mere 6% are recognized as high performers in terms of measurable earnings before interest and taxes (EBIT) impact, according to a 2025 report by McKinsey and Company. This discrepancy highlights that the challenge is not merely access to AI technology, but the organizational capacity to leverage it effectively. The release of advanced AI models like DeepSeek-R1 in early 2025 further democratized AI testing, shifting the strategic question from "Can we access AI?" to "Can our organization effectively utilize AI?"

This research delves into the internal mechanisms that differentiate successful AI adopters from those who struggle to realize value. It posits that digital orientation, a firm’s strategic commitment to integrating digital technologies, is not a direct driver of AI capability. Instead, its influence is mediated by organizational learning—the firm’s ability to acquire and apply new knowledge—and organizational forgetting—its capacity to shed outdated practices and assumptions. Furthermore, the study introduces Artificial Intelligence (AI) trust as a critical moderator, influencing how effectively digital orientation can be converted into these vital organizational processes.

Theoretical Framework: The Attention-Based View

At the heart of the study lies the attention-based view (ABV) of the firm, a theoretical lens that emphasizes how limited managerial attention shapes organizational decisions and actions. As proposed by Ocasio (1997), firms do not react to every opportunity or threat; rather, their behavior is dictated by the issues that decision-makers actively focus on. Strategic orientations, like digital orientation, are viewed as filters that direct this attention, prioritizing certain signals and allocating resources accordingly.

In this context, digital orientation is conceptualized as a firm-level pattern of attention that elevates digital and AI-related issues within managerial cognition. However, this attentional focus alone does not automatically translate into competence. The study argues that for digital orientation to foster AI capability, it must be channeled through organizational processes that absorb new knowledge and revise prior assumptions.

Dual Pathways to AI Capability: Learning and Forgetting

The research identifies two primary pathways through which digital orientation can cultivate AI capability:

• Organizational Learning: Drawing on established literature, organizational learning is defined as a firm’s capacity for experimentation, acquiring external knowledge, open communication, and employee involvement in problem-solving. A strong digital orientation is expected to stimulate these learning routines by making digital and AI concerns a consistent managerial priority and signaling that exploration is encouraged. This is crucial for AI applications, which often require iterative testing, interpretation of unexpected results, and continuous refinement of data and algorithms.

• Organizational Forgetting: The study also highlights the often-overlooked importance of organizational forgetting—the deliberate relinquishment of outdated routines, processes, and knowledge structures that hinder renewal. Unlike learning, which is additive, forgetting is subtractive. AI applications frequently challenge existing practices, such as replacing experience-based judgment with data-driven recommendations or shortening decision times with AI-enabled workflows. Without actively discarding old routines, firms may find AI tools used only superficially, or employees may resist integration. Digital orientation can signal the need for such shedding, but the firm’s ability to forget old practices is critical.

The Crucial Role of AI Trust

Recognizing that organizational members’ willingness to engage with AI is not guaranteed, the study introduces AI trust as a boundary condition. Defined as the degree to which managers and employees believe AI systems are reliable and useful for decision-making, AI trust is posited to strengthen both the learning and forgetting pathways. When AI trust is high, employees are more inclined to act on AI-driven insights and accept AI-based alternatives to familiar routines. Conversely, low AI trust can lead to superficial compliance, where AI outputs are reviewed but not genuinely relied upon, and existing routines remain a fallback.

Empirical Evidence and Key Findings

The study employed quantitative methods, collecting survey data from 306 middle and senior managers across diverse Chinese firms. Structural equation modeling (SEM) and fuzzy-set qualitative comparative analysis (fsQCA) were used to analyze the complex relationships.

The results confirmed that digital orientation has a positive direct association with AI capability. More significantly, both organizational learning and organizational forgetting were found to mediate this relationship, indicating that these internal processes are crucial conduits for translating strategic digital intent into practical AI competence. The indirect effect through organizational learning accounted for approximately 25% of the total effect, while the indirect effect through organizational forgetting explained about 22.5%. The direct effect of digital orientation on AI capability still held a substantial share at over 52%.

The moderating role of AI trust was also strongly supported. The study found that AI trust significantly strengthens the association between digital orientation and both organizational learning and organizational forgetting. This implies that a firm’s commitment to digital transformation is more likely to translate into effective learning and forgetting when employees and managers have confidence in the reliability and utility of AI systems. The indirect effects of digital orientation on AI capability were found to be significantly stronger at higher levels of AI trust.

Configurational Paths to High AI Capability

Complementing the SEM analysis, fsQCA identified distinct pathways or configurations that lead to high AI capability. The analysis revealed five sufficient pathways, with key conditions consistently appearing across these routes:

• Digital Orientation: This emerged as a core condition in all five pathways, underscoring its foundational importance.
• Organizational Forgetting: This condition was crucial in four of the five pathways, suggesting that the ability to shed outdated practices is often more critical than acquiring new knowledge for AI capability development.
• AI Trust: While not universally present in all pathways, AI trust was a core component in the "Trust-and-Forgetting Synergy" configuration, highlighting its activating role.

Interestingly, the analysis differentiated pathways based on firm maturity. Younger firms tended to follow a "trust-and-forgetting" route, where building AI trust and actively discarding old routines were paramount. Mature firms, on the other hand, often leveraged a "mature firm learning-driven" route, where established learning routines played a more dominant role.

Implications for Theory and Practice

The study offers significant theoretical contributions by:

1. Elucidating the Mechanism: It moves beyond treating digital orientation as a general predictor, framing it as an attentional configuration activated through learning and forgetting.
2. Integrating Knowledge Processes: It brings organizational learning and forgetting into a unified framework, demonstrating their complementary roles in AI capability formation.
3. Highlighting Trust’s Role: It identifies AI trust as a critical condition that governs whether digital orientation can activate the necessary knowledge processes.

For practitioners, the findings offer actionable insights:

• Translate Digital Orientation: Managers must ensure that stated digital strategies are translated into concrete organizational actions that foster AI experimentation, knowledge sharing, and capability development.
• Manage Learning and Forgetting Jointly: Firms need to cultivate both the capacity to acquire new digital knowledge and the courage to abandon outdated routines. These processes are not mutually exclusive but interdependent for effective AI integration.
• Cultivate AI Trust: AI trust should be treated as a strategic asset, deliberately nurtured through enhancing AI reliability, transparency, and governance to encourage employee engagement and reliance on AI systems.

The study’s authors, including W. Yang, Y. Ma, and Q. Yang, emphasized that AI capability development is a layered process, not solely dependent on technology or strategic pronouncements. It emerges from the interplay of strategic attention, knowledge renewal, and trust in AI. Future research is encouraged to explore this framework using longitudinal data, cross-country comparisons, and a more granular examination of AI capability dimensions to further refine our understanding of how organizations navigate the complex path to AI mastery.