
Architecting Agentic Systems
Engineering Dependable AI Agents
Buy the book
The full text is also free to read on this site.
How to cite this edition
Use the online edition at architecting-agentic-systems.net as the canonical reference. Copy a format below.
APA
Damian Beresford (2026). Architecting Agentic Systems: Engineering Dependable AI Agents (Online ed., v1.2.1). https://architecting-agentic-systems.net
Chicago (author-date)
Damian Beresford. Architecting Agentic Systems: Engineering Dependable AI Agents. Online ed., v1.2.1. 2026. https://architecting-agentic-systems.net.
BibTeX
@book{beresford2026agentic,
title = {Architecting Agentic Systems},
subtitle = {Engineering Dependable AI Agents},
author = {Damian Beresford},
year = {2026},
url = {https://architecting-agentic-systems.net},
note = {ISBN 9798185577516. Online edition, v1.2.1}
}
Copyright © 2026 Damian Beresford. All rights reserved.
Table of Contents
- Preface
- Chapter 1: The nature of agentic systems
- Chapter 2: The agent: Formal definition and properties
- Chapter 3: Architectural forces in agentic design
- Chapter 4: Cognitive patterns: A reference map
- Chapter 5: Bounded autonomy as discipline
- Chapter 6: Governance as architecture
- Chapter 7: Memory and state architecture
- Chapter 8: The ingestion pipeline: Architecting semantic memory
- Chapter 9: Control and coordination
- Chapter 10: The skills layer: Dynamic capability loading
- Chapter 11: Failure modes and anti-patterns
- Chapter 12: Testing, evaluation, and trace discipline
- Chapter 13: The glass layer: UI and interaction architecture
- Chapter 14: Enterprise SaaS integration
- Chapter 15: Model routing and edge infrastructure
- Chapter 16: System architectures
- Chapter 17: A worked example, Concord
- Chapter 18: Operationalization and production
- Chapter 19: The harness: architecting the agent loop
- Chapter 20: Glossary
- Chapter 21: Annotated bibliography
- Epilogue