Audit Quality Indicators (AQIs) have become vital tools for assessing and improving audit performance. This study examines how AQIs are defined, implemented, and interpreted across jurisdictions, with a particular focus on their integration into ESG (Environmental, Social, and Governance) assurance. Through a cross-jurisdictional comparison covering the EU, UK, US, and Canada, we analyze regulatory diversity and explore how different oversight models influence AQI adoption. Our findings reveal that AQIs play an increasingly important role not only in enhancing audit transparency but also in reinforcing the credibility of sustainability reporting under evolving frameworks like the EU’s Corporate Sustainability Reporting Directive (CSRD). However, implementation remains fragmented, and standardization challenges persist. The study offers empirical insights from a large-scale survey of audit professionals, highlighting how perceptions of audit quality vary by country, experience, and organizational context. We conclude that the harmonization of AQI frameworks, especially with ESG-focused metrics, is essential to foster public trust and ensure the accountability of sustainability assurance.

Offline Imitation Learning (IL), particularly through Behavioral Cloning (BC), offers a promising alternative to traditional reinforcement learning (RL) for inventory management, especially under data-scarce conditions such as cold-start scenarios. A notable limitation of BC relates to its sensitivity to covariate shift phenomenon, which leads to performance inconsistencies when the agents try to make decisions under “unseen” situations or distributions. In this paper we introduce a data-driven mechanism for data augmentation designed to mitigate covariate shift in inventory settings. The proposed method applies a geometric mirroring operator to discrete stateaction pairs, generating synthetic samples around empirical inventory thresholds. The cloning component is modeled by applying the established Multi-Layer Perceptron (MLP) architecture. Experimental evaluation on a real dataset demonstrates that the proposed mechanism significantly reduces KullbackLeibler (KL) divergence across multiple “unseen” state regions, outperforming both noise-injection and jitter-based baselines. These results underscore the efficacy of tailored augmentation for enhancing policy generalization in BC-based inventory systems and open new avenues for applying IL in managerial decisionmaking under uncertainty.

Background: The rapid evolution of omnichannel retailing has reshaped retail supply chains (SCs) by coupling replenishment, fulfillment, and service decisions across multiple demand channels under inventory, lead-time, and capacity constraints. These interdependencies create coordination challenges, particularly when demand shocks interact with limited operational capacity. Methods: To address these challenges, this study develops a centralized Hierarchical Reinforcement Learning (HRL) control framework that makes decision timing explicit: replenishment and allocation are optimized weekly, while fulfillment and lateral inventory rebalancing are controlled daily. Policies are learned using Proximal Policy Optimization (PPO) in an actor–critic architecture, with bounded stochastic policies for constrained action spaces. To mitigate the curse of dimensionality in HRL, we introduce a capacity-aware state–action encoding mechanism that compresses the control interface into structured summary signals. Demand shocks are modeled using two specifications: a mixed profile, where half the products follow a uniform demand process and the rest a Merton-type jump-diffusion process, and a fully shock-driven profile. Results: The framework is evaluated against forecast-driven base-stock and greedy fulfillment heuristics, and a perfect-information oracle, with pairwise differences examined through Wilcoxon signed-rank tests. Conclusions: Overall, the proposed framework improves learning efficiency and scalability, outperforming heuristic baselines while remaining below the oracle bound.