Librería Samer Atenea
Kálamo Books
Librería Elías (Asturias)
Librería Kolima (Madrid)
Librería Proteo (Málaga)
Agentic AI systems promise speed, intelligence, and autonomy-but without disciplined engineering, autonomy can turn small errors into costly failures.When Agents Fail: A Lifecycle Guide to Designing, Building, and Operating Agentic Systems That Fail Safely presents a practical, end-to-end framework for creating reliable, governable, and resilient agentic systems. From problem definition, data and schema design, and tool integration to testing, deployment, monitoring, incident response, continuous improvement, and safe retirement, the book treats autonomy as a lifecycle responsibility-not merely a technical feature.Real-world case studies illustrate how seemingly minor design weaknesses can produce cascading operational failures, and how observability, reversibility, accountability, and governance can prevent them.Designed for architects, engineers, technology leaders, and AI practitioners, this book provides the principles, practices, and operational discipline needed to build agentic systems that are not only capable-but trustworthy, recoverable, and safe when things go wrong.