intelligent automation

Production-grade autonomous AI workflows face significant engineering challenges in reliability, observability, maintainability, and security governance. Method: We propose a structured, full-lifecycle methodology comprising a multi-agent architecture with collaborative reasoning, tool augmentation, and dynamic orchestration—integrated with the Model Context Protocol (MCP), deterministic orchestration, pure function invocation, containerized deployment, and modular tool integration. We further define nine core engineering practices, including tool-first design, single-responsibility agents, externalized prompt management, and model-federation-driven responsible AI design. Contribution/Results: This work establishes the first systematic engineering paradigm for Agentic AI productionization, markedly improving system simplicity, observability, and governability. Empirical validation via a multimodal news analysis–media generation use case demonstrates robustness and scalability. The methodology provides a reusable framework and practical benchmark for industrial-scale autonomous AI systems.

Existing Robotic Process Automation (RPA) systems exhibit inherent limitations in symbolic processing, hindering the automation of complex, human-centric tasks. Method: Through a systematic literature review, this paper proposes the first taxonomy framework for intelligent RPA, structured around two meta-features—“integration” and “interaction”—and eight dimensions: architecture, capabilities, data, intelligence level, technical depth, deployment environment, lifecycle stage, and human–machine relationship. Contribution/Results: The framework transcends conventional RPA boundaries, explicitly characterizing diverse pathways for deep integration of machine learning and automation. It provides both theoretical foundations and practical guidance for expanding the scope of automatable tasks, thereby addressing a critical gap in the systematic classification of intelligent RPA systems.

Defining and monitoring complex robotic activities—comprising flexible, non-predefined sequences of atomic tasks—in dynamic, unstructured environments (e.g., precision agriculture) remains challenging. Method: This paper proposes an end-to-end framework integrating large language models (LLMs), automated planning, and symbolic reasoning. It enables humans to declaratively specify high-level procedures in natural language and supports explainable, real-time tracking and querying of task execution states across past, present, and future temporal scopes. Contribution/Results: To the best of our knowledge, this is the first approach realizing a closed-loop, natural-language-driven pipeline for high-level task planning and semantic-level monitoring—balancing adaptability with safety guarantees. Evaluated in real-world agricultural settings, the system accurately parses natural-language instructions, provides timely execution feedback, and significantly improves human-robot collaboration efficiency and process controllability.

This work addresses the absence of a unified conceptual framework for describing the autonomy of AI agents and the allocation of decision-making authority in contemporary CI/CD pipelines. It introduces the notion of “authority transfer” to systematically delineate the boundaries of agent autonomy, distinguishing between decision rights in the data plane and the control plane, and identifies governance of the control plane as a critical research direction. Through architectural abstraction, pattern identification, and governance mechanism design—supported by prototype implementation and analysis of industrial platforms—the study reveals three prevalent patterns: constrained autonomy, externally dominated governance, and delayed evaluation. These findings establish a theoretical foundation and outline a research agenda for developing safe, controllable, and highly autonomous CI/CD systems.

This work addresses the limitations of traditional industrial automation systems, which rely on fixed rules and struggle to autonomously interpret tasks or adapt in dynamic environments. The authors propose a novel three-layer framework that integrates large language models (LLMs) with digital twins, employing TPSR-based task modeling to translate natural language instructions into executable workflows. Four distinct LLM agent roles are designed to enable goal-directed, adaptive behavior. This study presents the first systematic integration of LLM agents and digital twins, endowing industrial systems with generalized reasoning and autonomous decision-making capabilities to support dynamic task planning and human–machine collaboration. Prototype experiments demonstrate high task executability, instruction fidelity, and generation accuracy, significantly reducing manual intervention while enhancing system adaptability and usability.

This study addresses the challenge of automating workflows in complex industries—such as logistics, healthcare, and construction—where processes are fragmented across heterogeneous tools and involve multi-party collaboration. The work proposes orchestration as a core abstraction to enable effective automation by dynamically coordinating multi-step tasks, enforcing domain-specific constraints, managing human approvals, and integrating legacy systems. It introduces the novel concept of “orchestration bottlenecks” and develops a theoretical framework that unifies multi-agent systems, workflow modeling, constraint reasoning, and human–AI collaboration, while exposing critical gaps in current multi-agent approaches at the orchestration level. Based on distinct sources of operational friction across domains, the paper advocates for targeted architectural safeguards—such as constraint enforcement or explainability—and phased implementation strategies to provide actionable pathways for automation in complex operational environments.

This work addresses the challenge of control-flow violations that arise when large language model (LLM) agents automate high-judgment quality management processes in regulated industries, often due to insufficient integration of symbolic structures such as regulatory rules and typed process models. To overcome this limitation, the paper introduces a “compliance-by-construction” paradigm, which internalizes compliance constraints as core components of the agent architecture rather than relying solely on external guardrails. By synergistically combining LLMs with symbolic systems—integrating typed process models, formal compliance constraints, and neuro-symbolic reasoning—the approach structurally prevents violations while preserving the ability to detect semantic errors. The study also systematically delineates the foundational and capability-level challenges required to realize this paradigm, offering a viable neuro-symbolic pathway for automation in regulation-intensive domains.

Enterprise operational workflows are notoriously difficult to automate end-to-end due to their heavy reliance on human intervention and limited adaptability to change. This work proposes the first action-centric workflow graph framework, which achieves automated construction, execution, and evolution through a three-stage pipeline: structured workflow graphs are extracted from human operation traces, executed via multi-agent online traversal, and continuously optimized in a closed loop using an Adaptive Traversal Reinforcement (ATR) mechanism. Integrating large-scale offline graph construction, graph-guided retrieval, and large language model reasoning, the approach was deployed across four cloud database services. It substantially outperforms the Trace-RAG baseline in coverage breadth, factual accuracy, and diagnostic throughput, achieving an expert blind-review score of 4.95 out of 5.

This study addresses the widespread lack of controllable, partially autonomous agentic AI systems tailored to the operational needs of small and medium-sized enterprises (SMEs), which hinders their ability to effectively reduce administrative burdens and enhance knowledge utilization. To bridge this gap, the paper proposes a human-centered, controlled Agentic AI integration framework designed for simple to moderately complex business processes. The framework ensures human oversight and accountability while enabling human-AI collaborative automation. It innovatively integrates dimensions of use-case adaptability, levels of autonomy, governance mechanisms, and employee empowerment, underpinned by core capabilities including goal interpretation, multi-step task planning, tool invocation, and seamless interaction with enterprise systems. Empirical evaluation demonstrates that the approach significantly improves process efficiency and organizational knowledge application, thereby validating Agentic AI as a viable and valuable productivity lever grounded in human-centric principles.

This work addresses the challenges industrial robots face in flexible automation—namely, accurately interpreting operator intent, verifying physical feasibility, and effectively recovering from execution failures—by proposing a human-robot collaborative neuro-symbolic framework. The approach integrates large language models for natural language understanding and contextual reasoning, augmented with an innovative SDI architecture inspired by the software engineering PGE pattern, LangGraph-based dynamic routing, and a two-tier recovery mechanism to enable precise structural-level replanning and geometric-level failure handling. A Unity3D digital twin further supports human-in-the-loop validation and correction. Experimental results demonstrate that the method significantly outperforms ten baseline approaches across natural language instructions of varying complexity, achieving the highest task success rate, while ablation studies confirm the effectiveness and necessity of each core component.