AI-powered self-healing enterprise applications: A new era of autonomous systems

 Corresponding author: Praveen Kumar Guguloth Copyright © 2025 Author(s) retain the copyright of this article. This article is published under the terms of the Creative Commons Attribution Liscense 4.0. AI-powered self-healing enterprise applications: A new era of autonomous systems Praveen Kumar Guguloth * State University of New York at Binghamton, USA. World Journal of Advanced Research and Reviews, 2025, 26(02), 754-761 Publication history: Received on 27 March 2025; revised on 03 May 2025; accepted on 05 May 2025 Article DOI: https://doi.org/10.30574/wjarr.2025.26.2.1682 Abstract This article introduces AI-powered self-healing enterprise applications as a transformative approach to maintaining system reliability and operational integrity. Traditional reactive maintenance strategies are increasingly inadequate in fast-paced digital environments where service interruptions directly impact business outcomes and customer loyalty. Self-healing systems represent a paradigm shift by leveraging artificial intelligence to detect issues proactively, diagnose root causes autonomously, and implement corrective measures without human intervention. The architecture of these systems encompasses monitoring layers, analysis engines, decision frameworks, execution modules, and knowledge repositories working in concert to maintain system health. Various integration patterns, including sidecar deployments, service meshes, orchestration frameworks, and embedded approaches, offer distinct advantages for different environments. Machine learning models and algorithmic techniques like time series analysis, clustering, natural language processing, classification, and causal inference enable sophisticated detection and remediation capabilities. Despite implementation challenges related to data quality, model drift, false positives, and organizational alignment, best practices have emerged to guide successful adoption. This article provides a comprehensive overview of selfhealing technologies and implementation strategies to help organizations achieve enhanced reliability in missioncritical enterprise applications. Keywords: Autonomous Remediation; AI-Driven Maintenance; Predictive Failure Detection; Operational Resilience; Enterprise Reliability 1. Introduction Enterprise applications form the backbone of modern business operations, making their reliability and availability paramount concerns for organizations worldwide. Reliability metrics such as Service Level Indicators (SLIs), Service Level Objectives (SLOs), and Service Level Agreements (SLAs) have become standard measurements for system performance, with 99.9% uptime (three nines reliability) allowing for approximately 8.76 hours of downtime per year [1]. Traditionally, system maintenance has relied on reactive strategies—responding to failures after they occur— resulting in significant downtime and business disruption. Error budgets, which define the acceptable threshold for system failures, typically range between 0.1% to 0.01% of total service time, translating to just minutes of allowable downtime per month for critical applications [1]. This approach has become increasingly inadequate in today's fast-paced digital economy where even minutes of service interruption can lead to substantial financial losses and damaged reputation. Research shows that downtime can cost businesses anywhere from $10,000 to $5 million per hour depending on the organization's size and industry sector [2]. For perspective, 98% of organizations report that a single hour of downtime costs over $100,000, while 81% indicate that 60 minutes of downtime impacts at least $300,000 in lost opportunity [2]. These figures don't account for the longterm impacts on customer trust and brand loyalty, with studies revealing that 91% of customers who experience service disruptions consider switching to competitors [2]. World Journal of Advanced Research and Reviews, 2025, 26(02), 754-761 755 The emergence of AI-powered self-healing systems represents a paradigm shift in how enterprise applications maintain operational integrity. These autonomous systems leverage sophisticated artificial intelligence techniques to detect issues before they impact users, diagnose root causes, and implement corrective measures without human intervention. By monitoring key reliability metrics like latency (request processing time), traffic (system load), errors (failed request rate), and saturation (system resource utilization)—collectively known as the LTES signals—self-healing systems can identify potential failures before they cascade [1]. Implementing these technologies has shown to reduce Mean Time To Detection (MTTD) by up to 60% and Mean Time To Resolution (MTTR) by approximately 43%, significantly improving the error budget utilization efficiency [1]. This evolution moves enterprise applications from simple fault tolerance to true operational resilience. When measuring reliability through Service Level Indicators (SLIs), organizations implementing comprehensive AI-powered self-healing frameworks have maintained 99.99% availability (four nines) compared to the industry standard of 99.9% (three nines), effectively reducing annual downtime from 8.76 hours to just 52.56 minutes [1]. Furthermore, automated remediation has demonstrated a 70% reduction in incidents that would typically require human intervention, allowing IT teams to focus on strategic initiatives rather than repetitive troubleshooting tasks [2]. This article examines the architecture, technologies, implementation challenges, and future trajectory of AI-powered self-healing enterprise applications, providing insights into how organizations can leverage these innovations to maintain competitive advantage in an increasingly digital marketplace. 2. Fundamental Architecture of Self-Healing Systems 2.1. Core Components Self-healing systems comprise several interconnected components that work in concert to maintain system health. Research on self-healing effectiveness metrics has demonstrated that systems implementing comprehensive monitoring and automated repair achieved a 65% success rate in addressing failures without human intervention, while partial implementations achieved only 42% success [3]. This difference becomes particularly significant in high-load conditions, where complete implementations maintain performance. The Monitoring Layer continuously collects performance metrics, logs, and operational data across the application stack. Studies showed that effective monitoring requires capturing both structural and behavioral properties, with systems tracking at least 12 distinct metrics achieving 22% higher anomaly detection rates [3]. Capturing state information at multiple abstraction levels proved crucial, with 3-tiered monitoring architectures demonstrating the highest efficiency in complex environments. The Analysis Engine processes collected data to identify anomalies, detect patterns, and predict potential failures. Experimental evaluation of self-healing systems revealed that rule-based analysis detected 76% of fault types, while machine learning models improved detection to 83% when sufficient training data was available [3]. The research demonstrated that multi-modal analysis approaches combining different techniques achieved the most comprehensive coverage. The Decision Framework determines appropriate remediation actions based on the analysis. Implementations using weighted decision trees achieved 44% faster recovery than simple if-then models [3]. Systems incorporating contextawareness into decision-making demonstrated a 37% improvement in selecting appropriate remediation strategies over context-free approaches. The Execution Module implements corrective measures automatically through orchestration frameworks. According to empirical measurements, automated recovery mechanisms successfully resolved 71% of detected failures [3]. Performance degradation during recovery averaged 18%, highlighting the importance of minimizing repair overhead. Finally, the Knowledge Repository maintains a database of historical incidents, successful remediation strategies, and system behavior patterns. Case-based reasoning approaches leveraging historical data improved remediation success rates by 28% compared to static rule-based systems [3]. World Journal of Advanced Research and Reviews, 2025, 26(02), 754-761 756 2.2. Integration Patterns Self-healing capabilities can be integrated into enterprise applications through various architectural patterns. Research on autonomous remediation strategies has demonstrated distinct effectiveness profiles for different integration approaches. The Sidecar Pattern deploys monitoring and self-healing capabilities alongside application containers as companion processes. Testing of autonomous remediation agents showed a 76.5% effectiveness rate when deployed as sidecars with minimal performance impact of 4-7% overhead [4]. This pattern provided isolation between remediation logic and application code, reducing potential failures by 31%. Service Mesh implementations provide network-level self-healing infrastructure that manages service-to-service communication. Experimental evaluations demonstrated that service meshes intercepting anomalous traffic patterns achieved 82% mitigation effectiveness for network-related issues [4]. Automatic retry policies with exponential backoff reduced system-wide impact by 66% during partial outages. Orchestration Frameworks offer built-in self-healing capabilities through health probes and automatic pod replacement. Studies of autonomous remediation in containerized environments showed 89% effectiveness in addressing infrastructure-level failures [4]. Recovery time measurements averaged 31.5 seconds, significantly outperforming manual intervention times of 10.2 minutes. The Embedded Approach integrates self-healing logic directly into application code through resilience libraries. Instrumented applications demonstrated 68% effectiveness in self-correction while adding approximately 12% code complexity [4]. This approach showed particular strength in handling application-specific anomalies that infrastructure-level healing could not detect. Table 1 Comparative Effectiveness of Self-Healing Implementation Approaches [3,4] Implementation Approach Effectiveness Rate (%) Orchestration Frameworks 89.0 Machine Learning Models 83.0 Service Mesh 82.0 Sidecar Pattern 76.5 Rule-based Analysis 76.0 3. AI Technologies Powering Self-Healing Mechanisms 3.1. Machine Learning Models Various machine learning approaches underpin modern self-healing systems, each contributing unique capabilities to autonomous remediation frameworks. According to recent research, self-healing systems implementing supervised learning models achieve fault prediction accuracy rates of 82% when trained on properly labeled historical incident data, providing critical early warning for potential system failures [5]. These models analyze patterns in system behavior, with the most effective implementations collecting at least 14 days of historical metrics to establish baseline performance parameters. Unsupervised learning techniques have demonstrated significant value in identifying anomalous system behavior without prior examples of failures. Studies show that clustering-based anomaly detection algorithms can identify up to 78% of novel failure modes that would otherwise go undetected by traditional rule-based monitoring systems [5]. Production implementations have shown that these models require approximately 40% less maintenance effort compared to manually configured alerting thresholds, which typically need adjustment every 2-3 months as application behavior evolves. Reinforcement learning approaches improve recovery strategies over time by evaluating remediation action success. Analysis of real-world implementations indicates that RL-based self-healing systems improve their remediation success rates by approximately 15% over the first six months of operation [5]. The most effective implementations utilize World Journal of Advanced Research and Reviews, 2025, 26(02), 754-761 757 reward functions that balance multiple objectives, with 60% of the scoring based on time-to-recovery metrics and 40% based on minimizing service disruption during remediation actions. Deep learning models process complex telemetry data to detect subtle indicators of impending failures. Research shows that convolutional neural networks applied to system metrics can identify precursor patterns for 73% of major incidents with an average lead time of 27 minutes before service impact [5]. Production implementations typically require training on a minimum of 200 labeled incidents to achieve reliable results, with transfer learning approaches reducing this requirement by up to 40% for similar system architectures. 3.2. Key Algorithmic Techniques Time series analysis techniques form the foundation of many self-healing systems, with research indicating that sophisticated forecasting models like Prophet can achieve 91% accuracy in predicting resource utilization anomalies when trained on a minimum of 30 days of historical data [5]. Implementations utilizing these techniques have demonstrated the ability to identify potential failures up to 45 minutes before traditional threshold-based monitoring systems trigger alerts. Clustering algorithms enable efficient incident categorization, with k-means clustering demonstrating 83% accuracy in identifying distinct failure categories across heterogeneous infrastructure components [5]. Systems employing these techniques have shown a 62% reduction in mean time to repair by quickly matching current incidents with previously resolved cases that share similar characteristics. Natural language processing plays a critical role in autonomous data healing, with recent research showing that transformer-based models can achieve 86% accuracy in identifying data integrity issues from unstructured log files [6]. These models can process approximately 10,000 log entries per minute, extracting actionable information with 79% precision and 74% recall rates across diverse logging formats. Classification models are essential for incident prioritization, with gradient-boosted decision trees demonstrating 88% accuracy in determining incident severity across a sample of 12,000 historical events [6]. Autonomous data healing systems implementing these models have achieved a 31% reduction in critical data integrity incidents by correctly prioritizing preventive actions based on predicted impact. Causal inference models determine root causes by establishing relationships between observed symptoms and underlying issues. Graph-based approaches have shown 77% accuracy in identifying the true source of data integrity problems in complex relational databases, analyzing up to 500 table relationships simultaneously [6]. These models reduce diagnostic time by an average of 47 minutes per incident compared to manual investigation approaches. Table 2 Performance Analysis of AI Techniques for Autonomous Remediation [5,6] AI Technology Accuracy/Effectiveness Rate (%) Time Series Analysis (Prophet) 91.0 Classification Models (Gradient-Boosted Trees) 88.0 Natural Language Processing (Transformer-based) 86.0 Clustering Algorithms (k-means) 83.0 Supervised Learning Models 82.0 4. Real-World Implementation Scenarios 4.1. Cloud Infrastructure Self-Healing Cloud-based applications leverage AI-driven self-healing to maintain high availability. Studies of quantum-enhanced optimization in self-healing cloud systems demonstrate a 67% reduction in mean time to recovery compared to classical approaches, with recovery times decreasing from an average of 17 minutes to just 5.6 minutes [7]. This significant improvement directly contributes to enhanced service availability, with measured uptime increasing from 99.91% to 99.97% across studied implementations. World Journal of Advanced Research and Reviews, 2025, 26(02), 754-761 758 Resource Optimization mechanisms automatically scale infrastructure based on demand predictions, with quantumenhanced forecasting models showing 83% accuracy in predicting resource requirements up to 22 minutes in advance [7]. This predictive capacity enables precise scaling that reduces resource over-provisioning by 28% while simultaneously decreasing performance degradation incidents by 52%, resulting in optimal resource utilization. Automated Failover systems initiate instance migration when hardware failures are predicted, with quantum-enhanced detection algorithms identifying 75% of imminent failures approximately 8 minutes before occurrence [7]. This early detection enables proactive workload migration that preserves system state and user sessions, reducing average downtime per incident by 84% compared to traditional reactive approaches. Configuration Drift Detection identifies and corrects unauthorized or problematic configuration changes, with machine learning models capable of detecting 89% of potentially harmful configuration drift within 3.7 minutes of occurrence [8]. These systems automatically remediate 63% of identified issues without human intervention, significantly reducing the window of vulnerability and preventing escalation to service-impacting incidents. Network Performance Optimization reroutes traffic when congestion or latency issues are detected, with AI-driven routing algorithms reducing average response latency by 47% during peak traffic periods [8]. These systems identify optimal routing paths with 82% accuracy, implementing traffic adjustments an average of 7 minutes before traditional threshold-based alerts would trigger manual intervention. 4.2. Database and Storage Systems Database systems benefit significantly from self-healing capabilities. Research across production environments shows implementation of intelligent monitoring reduced unplanned database downtime by 65% while improving overall query performance by 37% [7]. These improvements translate to substantial operational efficiency gains and enhanced user experience. Query Performance Tuning mechanisms automatically optimize slow-running queries, with quantum-enhanced analysis identifying optimization opportunities for 78% of problematic queries [7]. The autonomous implementation of these optimizations results in an average execution time improvement of 54%, with complex analytical queries showing the most dramatic improvements of up to 72% reduced execution time. Index Management creates, rebuilds, or reorganizes indexes based on usage patterns, with machine learning models identifying optimal indexing strategies with 85% accuracy [8]. Automated implementation of these recommendations reduces index fragmentation by 61%, translating to a 33% improvement in query throughput across common workloads. Storage Allocation mechanisms preemptively allocate additional storage before capacity limits are reached, with forecasting models demonstrating 90% accuracy in predicting storage requirements up to 9 days in advance [7]. This predictive capacity enables proactive resource allocation that prevents 97% of potential storage-related outages. Data Corruption Prevention detects and addresses potential corruption issues before they propagate, with pattern recognition algorithms identifying 83% of corruption signatures before data integrity is compromised [8]. Early detection enables successful remediation in 71% of cases without data loss, significantly improving recovery outcomes compared to traditional reactive approaches. 4.3. Application-Level Self-Healing Within application code, self-healing mechanisms provide resilience. Research across production deployments shows applications implementing comprehensive self-healing architectures experience 68% fewer critical failures and recover from unavoidable incidents 62% faster than traditional implementations [8]. Memory Leak Detection identifies and addresses memory management issues before they cause crashes, with machine learning models successfully detecting 91% of memory leaks an average of 43 minutes before application failure [8]. Autonomous remediation successfully resolves 74% of these issues through techniques like selective object cleanup and targeted service restart. Deadlock Resolution automatically detects and breaks deadlocks in transaction systems, with graph-based analysis identifying circular dependencies with 89% accuracy [7]. Self-healing mechanisms successfully resolve 72% of potential deadlocks while preserving data integrity, dramatically reducing transaction timeouts in production environments. World Journal of Advanced Research and Reviews, 2025, 26(02), 754-761 759 API Dependency Management implements circuit breakers and fallback mechanisms for external service dependencies, maintaining 83% of critical functionality during dependency failures [8]. These systems dynamically adjust failure thresholds based on observed patterns, reducing cascading failures by 67% compared to static configurations. Session Management preserves user session data during backend service transitions, with distributed caching mechanisms successfully maintaining 87% of active sessions during infrastructure failures [7]. These approaches reduce average service interruption from 35 seconds to just 4 seconds during backend transitions, preserving user experience during maintenance events. Table 3 Effectiveness Comparison of Self-Healing Technologies in Production Environments [7,8] Implementation Area Improvement Rate (%) Storage Outage Prevention 97.0 Memory Leak Detection 91.0 Configuration Drift Detection 89.0 Session Preservation During Failures 87.0 Resource Requirement Prediction 83.0 5. Implementation Challenges and Best Practices 5.1. Technical Challenges Organizations implementing self-healing systems face several significant hurdles that can impact effectiveness. Data quality issues represent a fundamental challenge, with insufficient or low-quality monitoring data hampering effective analysis. According to industry research, organizations typically monitor only 30% of their IT infrastructure effectively, leaving significant blind spots that prevent comprehensive self-healing capabilities [9]. This gap in observability directly affects detection capabilities, with incomplete monitoring coverage reducing incident detection rates by up to 45%. Model drift presents a persistent challenge as AI models become less effective as application behavior changes over time. Studies show that without regular maintenance, AI model effectiveness decreases by approximately 25% annually as application architectures and usage patterns evolve [9]. This degradation requires teams to implement continuous model retraining and validation procedures to maintain detection accuracy above acceptable thresholds. False positives emerge when overzealous systems implement unnecessary remediation actions, creating operational disruptions. Initial implementations typically experience false positive rates between 10-15%, potentially causing more disruption than the issues they aim to solve [9]. Establishing proper baseline behavior and implementing progressive confidence thresholds can reduce these rates to under 5% during the first six months of operation. Complexity management challenges arise as self-healing systems add another layer of sophistication to already complex enterprise applications. Research indicates that 78% of organizations underestimate the integration complexity of autonomous systems, leading to implementation delays averaging 3-4 months longer than initially projected [9]. 5.2. Organizational Considerations Beyond technical aspects, successful implementation requires organizational alignment. The skills gap presents a substantial hurdle, as teams need expertise in both AI and traditional operations to maintain self-healing systems. Research across multiple industry sectors indicates that 72% of organizations report significant skills gaps when implementing advanced automation technologies, with only 25% having developed comprehensive upskilling programs to address these deficiencies [10]. Trust building represents a critical organizational consideration, as stakeholders must develop confidence in autonomous systems making critical decisions. Studies show that approximately 65% of stakeholders initially express reservations about automated decision-making in critical infrastructure, with trust developing progressively as systems demonstrate reliability [10]. Organizations reporting successful implementations typically demonstrate a structured approach to building confidence through transparent operations and clear communication. World Journal of Advanced Research and Reviews, 2025, 26(02), 754-761 760 A hybrid approach combining human oversight with automated remediation provides a balanced solution that addresses organizational concerns. Research indicates that 83% of successful implementations utilize a tiered autonomy model where routine issues are fully automated while complex scenarios maintain human oversight [9]. This balanced approach typically reduces incident resolution times by 60-70% while maintaining appropriate governance. Change management challenges emerge when shifting from reactive to predictive operations, requiring cultural adaptation. According to organizational readiness research, only 32% of organizations adequately prepare their teams for the significant workflow changes introduced by autonomous systems [10]. Successful transitions typically involve all key stakeholders from early design phases, with approximately 15-20% of project resources dedicated specifically to change management activities. 5.3. Best Practices Starting small by beginning with non-critical components before expanding to mission-critical systems significantly increases success rates. Organizations implementing an incremental approach report 70% higher satisfaction with outcomes compared to those attempting comprehensive deployments [9]. Beginning with systems that have clear failure modes and minimal cross-dependencies provides valuable learning opportunities while limiting potential negative impacts. Comprehensive monitoring established before implementing automated remediation provides a solid foundation. Research indicates that organizations investing in monitoring infrastructure for at least 3-4 months before enabling automated remediation experience 40% fewer implementation issues [9]. This preparatory phase ensures sufficient data quality and coverage for effective anomaly detection and root cause analysis. Human-in-the-loop design incorporating approval workflows for high-impact remediation actions balances automation with appropriate oversight. Studies of organizational readiness for advanced automation indicate that 78% of successful implementations maintain human oversight for critical systems, particularly during initial deployment phases [10]. This approach builds stakeholder confidence while providing a safeguard against potential automation errors. Continuous learning mechanisms implement feedback loops to improve AI model performance over time. Research shows that organizations implementing structured feedback processes achieve approximately 30% higher model accuracy after six months compared to static deployments [10]. This improvement directly correlates with reduced false positives and higher stakeholder confidence in system recommendations. Documentation maintaining records of all autonomous actions enables effective audit and analysis. Organizations implementing comprehensive action logging report approximately 45% faster troubleshooting for complex incidents by providing clear visibility into system behavior and decision rationale [9]. Table 4 Critical Factors Affecting Self-Healing System Success Rates [9,10] Challenge Area Impact Rate (%) Underestimated Integration Complexity 78.0 Skills Gap in Organizations 72.0 Initial Stakeholder Reservation 65.0 Reduction in Detection Capabilities 45.0 Annual Model Effectiveness Degradation 25.0 6. Conclusion AI-powered self-healing enterprise applications represent a fundamental evolution in system reliability and operational efficiency. By shifting from reactive to proactive maintenance paradigms, organizations can significantly reduce downtime, lower operational costs, and improve user experiences. The journey toward fully autonomous self-healing systems continues to advance with developments in machine learning, edge computing, and causal AI promising increasingly sophisticated capabilities. As these technologies mature, self-healing will likely become a standard feature rather than a competitive advantage. Organizations embracing this technology now will benefit from improved reliability while gaining valuable experience in managing AI-driven autonomous systems—expertise that will prove World Journal of Advanced Research and Reviews, 2025, 26(02), 754-761 761 increasingly valuable as AI transforms enterprise technology landscapes. The future of enterprise applications lies not just in functionality and performance but in resilience and autonomy, with self-healing systems leading this transformation by creating applications that actively maintain their operational integrity with minimal human intervention. References [1] Abdulsalaam Noibi, "10 Essential Reliability Metrics for Software Quality," SigNoz, 2024. [Online]. Available: https://signoz.io/guides/reliability-metrics/ [2] Orion Network Solutions, "How Downtime With Information Systems Can Cost Business Thousands In Lost Opportunity," Orionnetworks.net. [Online]. Available: https://www.orionnetworks.net/how-downtime-withinformation-systems-can-cost-business-thousands-in-lost-opportunity/. [3] Aaron B. Brown and Charlie Redlin, "Measuring the Effectiveness of Self-Healing Autonomic Systems," Proceedings of the Second International Conference on Autonomic Computing (ICAC’05). [Online]. Available: https://www.netlab.tkk.fi/opetus/s384030/k06/papers/MeasuringTheEffectivenessOfSelfHealing.pdf [4] Jabir Abbas Sambo, "Developing Autonomous Remediation Strategies for Network Security," Global Scientific Journal, vol. 12, Issue 7, 2024. [Online]. Available: https://www.globalscientificjournal.com/researchpaper/DEVELOPING_AUTONOMOUS_REMEDIATION_STRAT EGIES_FOR_NETWORK_SECURITY_.pdf [5] Karthigayan Devan, "Building Self-healing Systems Using AI And Machine Learning: Advanced Platform Engineering Practices," International Journal of Innovation Studies, 2023. [Online]. Available: www.researchgate.net/publication/389098378_Building_Selfhealing_Systems_Using_AI_And_Machine_Learning_Advanced_Phttps://latform_Engineering_Practices [6] Babita Kumari, "Autonomous Data Healing: AI-Driven Solutions for Enterprise Data Integrity," Autonomous Data Healing: AI-Driven Solutions for Enterprise Data Integrity, 2025. [Online]. Available: https://www.researchgate.net/publication/387880145_Autonomous_Data_Healing_AIDriven_Solutions_for_Enterprise_Data_Integrity [7] Mahender Singh, "Self-Healing Cloud Infrastructures via Al-Driven Quantum Optimization," ResearchGate, 2024. [Online]. Available: https://www.researchgate.net/publication/390827738_SelfHealing_Cloud_Infrastructures_via_Al-Driven_Quantum_Optimization [8] Habeeb Agoro, "Building Resilient Software Systems: SelfHealing Architectures with Machine Learning," ResearchGate, 2022. [Online]. Available: https://www.researchgate.net/publication/390768408_Building_Resilient_Software_Systems_Self_Healing_Architectures_with_Machine_Learning [9] Derek Pascarella, "Future-Proof Your IT: Understanding Self-Healing IT Infrastructure," Resolve, 2025. [Online]. Available: https://resolve.io/blog/guide-to-self-healing-it-infrastructure [10] Jiju Antony et al., "An exploration of organizational readiness factors for Quality 4.0: an intercontinental study and future research directions," International Journal of Quality & Reliability Management ahead-of-print(aheadof-print), 2022. [Online]. Available: https://www.researchgate.net/publication/357510099_An_exploration_of_organizational_readiness_factors_fo r_Quality_40_an_intercontinental_study_and_future_research_directions