Agentic AI and Autonomous Systems

Агентный искусственный интеллект и автономные системы
S. Prince, K DanielRaj, A. Anu Priya
2026-06-24

Agentic artificial intelligenceAutonomous systemsLarge language model-powered agentsMulti-agent coordinationRetrieval-Augmented Generation
This chapter examines the rapidly evolving paradigm of Agentic Artificial Intelligence and Autonomous Systems, investigating how large language model-powered agents transition from passive query-response tools to active, goal-directed entities capable of multi-step reasoning, tool use, memory management, and self-correction. This chapter per the authors presents a formal taxonomy of agent architectures alongside analysis of enabling frameworks including ReAct, Reflexion, LangGraph, AutoGen, and CrewAI. Core topics include multi agent coordination, Retrieval-Augmented Generation, the Model Context Protocol, neuromorphic computing, and multimodal perception for embodied AI. Case studies spanning autonomous software engineering, clinical decision support, financial trading, and self-driving vehicles are evaluated. The chapter concludes with safety, alignment, governance, and quantum computing trajectories that will amplify agentic AI capabilities,contributing an engineering-first framework essential for researchers, practitioners, and policymakers navigating the Quantum AI era.
1
Agentic AI systems advance beyond passive question answering by supporting goal-directed, multi-step reasoning, tool use, memory management, and self-correction.
2
Case studies examine applications of autonomous agents in software engineering, clinical decision support, financial trading, and self-driving vehicles.
3
Multi-agent coordination, Retrieval-Augmented Generation, the Model Context Protocol, neuromorphic computing, and multimodal perception are identified as key enabling technologies.
4
The chapter emphasizes that safety, alignment, governance, and emerging quantum-computing trajectories are central to the future development of agentic AI.
5
The chapter presents a formal taxonomy of agent architectures and analyzes frameworks including ReAct, Reflexion, LangGraph, AutoGen, and CrewAI.

large language model-powered agents and autonomous systems

their transition from passive query-response tools to active, goal-directed entities, including multi-step reasoning, tool use, memory management, self-correction, coordination, perception, and safety

Publication Details
Publication Date
2026-06-24
Journal
Publisher
ISSN
Cited by
2
Access Type
Author Information
Authors
S. Prince
K DanielRaj
A. Anu Priya
Explore further
Open the scid.ai AI chat with a ready-made request: it will find papers on a similar topic and help build a literature review.
Find similar papers in the chat →
Make a presentation
100%