AI readiness

AI readiness is the organizational capacity to make a specific AI adoption purpose viable. 2021-johnk-ready-or-not-ai-comes makes the construct purpose- and context-specific: readiness is not a generic maturity score or a one-time gate before adoption. Because AI is a general-purpose technology, organizations have to assess readiness against the particular use case, work process, customer interface, and adoption stage they are trying to move through.

Jöhnk, Weißert, and Wyrtki organize readiness into five categories: strategic alignment, resources, knowledge, culture, and data. The factors are concrete enough to be useful diagnostically — AI-business potentials, customer AI readiness, top-management support, AI-process fit, data-driven decision-making, financial budget, personnel, IT infrastructure, AI awareness, upskilling, AI ethics, innovativeness, collaborative work, change management, data availability, data quality, data accessibility, and data flow — but the paper treats the indicators as illustrative rather than validated measures.

The important move for this wiki is that readiness and adoption are mutually reinforcing. Readiness makes successful adoption more likely, but successful smaller-scale adoption also builds readiness by producing experience, acceptance, and clearer understanding of where AI can create value. That makes readiness a dynamic design problem, not a static compliance checklist.

2026-bakker-artificial-intelligence-job-design adds a JD-R distinction that keeps this topic from treating readiness as solely enterprise infrastructure. It separates AI-readiness job resources—training, support, time and autonomy to experiment, psychological safety, responsibility clarity, and supportive leadership—from AI-readiness personal resources such as self-efficacy, digital literacy, metacognitive skill, domain expertise, and learning orientation. The practical implication is a readiness-fit test: mandated AI in a low-resource job is more likely to produce strain, disengagement, resistance, or unsafe workarounds, while skill-building alone will not repair missing time, support, participation, or control. These are theory-led categories rather than validated readiness scales.

2026-reich-work-design-multidimensional-ai-threat adds a job-design boundary to readiness. Even if organizational knowledge, culture, and resources are present, adoption may still vary by whether roles provide autonomy and skill variety for experimentation. Readiness therefore has a work-design layer below the organization-level checklist.

2025-tona-deployment-ai-infer-employee-skills adds a workforce-capability layer to readiness. J&J’s case treats digital-transformation readiness as partly a problem of knowing which skills the current workforce has, which future capabilities the organization values, and whether AI-inferred skill data can support reskilling and internal-mobility decisions. Readiness therefore includes the measurement and governance capacity to make workforce capability visible without turning employee data into an unvalidated or untrusted control system.

2025-hautamaki-fully-leveraging-ai-b2b-sales adds a managerial-capability layer to readiness. Its grounded-theory framework says B2B sales managers need data-based human capital, social capital for knowledge-sharing culture, and a transformative AI-positive mindset to convert available AI tools into individual knowledge creation, team knowledge transfer, and organizational knowledge-asset reconfiguration. Readiness therefore includes whether middle and frontline managers can translate AI into domain-specific knowledge processes, not only whether the organization has strategy, data, infrastructure, or top-management support.

2025-schweitzer-leading-ai-adoption adds a behavioral-change-management layer to readiness. If leaders treat AI implementation as an engineering handoff, readiness claims are incomplete even when strategy, resources, knowledge, culture, and data look adequate. Organizations also need the behavioral capability to anticipate designer bias, employee resistance, algorithm aversion, autonomy threat, leader overconfidence, and escalation of commitment across design, adoption, and management stages.

2018-fugate-change-compliance-championing adds a more specific change-support readiness mechanism: employees are more likely to appraise change as a challenge when management provides useful information, clear mutual expectations, and trustworthy change competence. Readiness for AI transformation therefore includes whether leaders can produce the conditions for employee championing, not only whether the organization has AI strategy, resources, data, and skills.

1997-goh-benchmarking-learning-capability adds a pre-AI organizational-learning layer to readiness. Its five learning-capability conditions — shared purpose, leadership commitment and empowerment, experimentation and rewards, knowledge transfer, and teamwork/group problem solving — name the infrastructure organizations need if AI pilots are going to become changed routines rather than isolated experiments. Readiness therefore includes whether the organization can learn from AI experimentation, failures, external examples, and local practice variation.

2026-shin-effective-cognitive-population applies a narrower readiness logic at national scale. Its ECP planning ledger separates HCI+-based human productivity from AI opportunity (digital infrastructure plus innovation/economic integration) and conversion governance (regulation and ethics), then makes the combination and its trade-offs inspectable. It does not validate a generic readiness score or show that preparedness causes output; instead, it supports the more limited proposition that readiness claims should specify the capability, opportunity, and governance conditions they combine and avoid double-counting them.

2026-wang-whose-readiness-counts adds a reporting rule for readiness assessments: a sector score can help scan a portfolio, but it should not stand in for agreement about any particular application. In the study, differences between broad challenge-family averages explained only about 2% of the observed variation in perceived preparedness. Reports therefore need to retain application-level results, show how much people disagreed, and disclose whose judgments produced the average. This is evidence about perceived community preparedness, not an audit of an organization’s actual capability.

Connections

• Connects to ai-adoption because readiness is one way of naming what binds adoption after tools become available: not access, but organizational capability, commitment, data, and fit for the intended adoption purpose.
• Connects to work-redesign through AI-process fit, data flow, and the need to reshape work processes before AI systems can produce reliable value.
• Connects to ai-mediated-work-experience because readiness is partly experienced through autonomy, upskilling, workload, identity threat, and local support for experimentation.
• Connects to construct-validity if readiness factors become survey or assessment instruments: the paper proposes factors and indicators, but future work still has to validate whether those indicators coherently measure readiness.
• Connects to human-ai-collaboration through customer/user readiness, upskilling, AI awareness, and AI ethics — adoption quality depends on humans understanding, accepting, and governing AI-enabled work.
• Connects to competency-modeling because readiness for digital transformation may require a future-oriented skills architecture before upskilling, mobility, or AI deployment can be targeted coherently.
• Connects to ai-mediated-organizational-networks because readiness depends on whether managers can build the cross-functional ties through which AI use cases, data knowledge, and domain expertise move across teams.
• Connects to behavioral-human-centered-ai because readiness includes the capacity to design adoption around real human behavior rather than rational-user assumptions.
• Connects to employee-change-support because readiness includes credible change management support: information, mutual expectation clarity, and trust in management competence during disruption.
• Connects to organizational-learning-capability because AI readiness depends on whether experimentation, knowledge transfer, and team problem solving can turn local AI use into shared organizational practice.
• Connects to ai-as-a-job-resource because AI becomes usable support only when the job and personal resources needed to interpret, verify, integrate, and learn from it are present.

Contradictions & open questions

• 2021-johnk-ready-or-not-ai-comes frames readiness broadly across organizational factors, while 2026-kim-mapping-ai-into-production isolates a narrower causal bottleneck: managerial discovery of where AI creates value. The relationship between broad readiness assessments and high-leverage mapping interventions is open.
• The proposed 58 indicators are illustrative, not psychometrically validated. If AI readiness becomes an organizational assessment construct, the measurement work still has to establish dimensionality, reliability, criterion validity, and incremental validity over existing innovation/adoption readiness constructs.
• Readiness may be partly endogenous to adoption. If successful pilots build readiness, then cross-sectional readiness scores could confuse preconditions with outcomes unless timing is modeled carefully.
• 2026-reich-work-design-multidimensional-ai-threat raises the same timing problem at the job level: autonomy and skill variety may enable AI adoption, be perceived differently by adopters, or both.
• 2025-tona-deployment-ai-infer-employee-skills raises a measurement-governance tension: AI skills inference can make readiness more visible, but readiness claims become fragile if inferred skills are treated as more accurate, complete, or employee-accepted than the evidence supports.
• 2025-hautamaki-fully-leveraging-ai-b2b-sales raises a level-of-analysis tension: readiness can be present at the enterprise level while AI exploitation still depends on sales managers’ local capabilities and team knowledge routines. The framework needs quantitative and field-performance evidence before those capabilities can be treated as readiness dimensions rather than promising qualitative categories.
• 2025-schweitzer-leading-ai-adoption raises a construct-boundary tension: behavioral change-management capability could be part of AI readiness, a method for producing readiness, or a post-adoption capability built through experience. 2018-fugate-change-compliance-championing sharpens the issue by showing that change management support predicts appraisals and support behavior, but it still leaves open whether employee championing should be measured as a readiness antecedent, an implementation process, or an adoption outcome in AI transformation.
• 1997-goh-benchmarking-learning-capability raises a timing problem for readiness: learning capability may be an antecedent to AI adoption, an outcome of repeated AI experimentation, or both. Cross-sectional readiness surveys need to avoid treating learned implementation capacity as if it were purely pre-existing maturity.
• 2026-shin-effective-cognitive-population finds that its national opportunity-governance block improves concurrent output fit but does not support a direct opportunity × governance interaction. The unresolved issue is whether a conjunctive readiness rule is a useful precautionary planning constraint, a descriptive index choice, or a causal proposition that needs longitudinal and multilevel evidence before it can guide intervention.
• 2026-wang-whose-readiness-counts shows that broad readiness averages can conceal disagreement among people and applications, but its survey does not establish which level of disagreement should block, slow, or reshape a real deployment. The practical threshold for treating a readiness score as adequate evidence remains open.
• 2026-bakker-artificial-intelligence-job-design gives AI-readiness resources a coherent JD-R role, but its proposed job-resource and personal-resource categories overlap conceptually with existing organizational readiness, AI literacy, self-efficacy, learning orientation, and supportive-climate constructs. The wiki still needs discriminant-validity and intervention evidence showing whether the two-level distinction predicts adoption quality, well-being, or performance beyond those adjacent measures.