No. 262 of 500 No. 30 in Generative AI & LLMs

Agentic Design Patterns

A Hands-On Guide to Building Intelligent Systems

By Antonio Gullí

Print length
470 pages
Publisher
Springer
Reading level
Intermediate
ISBN-13
978-3032014016
Published
October 31, 2025
Written for
Software engineers and technical architects building LLM-based agents who want a structured pattern vocabulary.

Synopsis

Antonio Gullí, a senior director in Google's Office of the CTO, organizes the craft of building AI agents into 21 design patterns, one per chapter, in the spirit of the classic software design-patterns books. The sequence starts with fundamentals such as prompt chaining, routing, parallelization, reflection, tool use, planning and multi-agent collaboration, then moves to memory management, learning and adaptation, the Model Context Protocol and goal monitoring. Later parts cover exception handling, human-in-the-loop review, retrieval-augmented generation, agent-to-agent communication, resource-aware optimization, reasoning techniques, guardrails, evaluation, prioritization and exploration. Appendices add advanced prompting techniques and a survey of agent frameworks.

Each chapter follows a fixed template: an overview of the pattern, practical use cases, runnable code and a summary of key takeaways. Examples rotate among LangChain and LangGraph, CrewAI and Google's Agent Development Kit, so readers see how one idea maps onto different toolkits. Springer published the book in October 2025, and Gullí pledges all author royalties to Save the Children. Readers gain a common language for designing and reviewing agent architectures and a reference to consult when a system needs, say, a reflection loop or a guardrail layer. The code, however, is tied to fast-moving libraries and may need updating.

Scorecard

Editorial score

How it scored

Editorial score: 72.8 / 1005 weighted criteria

  • Influence30%

    19.5/30
  • Authority25%

    18.8/25
  • Practical value20%

    16.0/20
  • Currency15%

    12.0/15
  • Readability10%

    6.5/10
  • Editorial score

    72.8/100

How the rank was computed

Rank No. 262 of 500 comes from a composite of 54.3: observable signals 21.6 of 55 points plus editorial review 32.7 of 45 points.

All 11 inputs
InputPoints
Observable signals21.6/55
Reach (Amazon ratings volume)7.5/16.5
Consensus across best-of lists5.5/13.8
Awards and bestseller evidence0.0/8.3
Reader rating, sample-adjusted4.3/8.3
Amazon sales rank1.5/5.5
Recency2.8/2.8
Editorial review32.7/45
Author authority8.4/11.3
Influence8.8/13.5
Practical value7.2/9.0
Readability2.9/4.5
Currency5.4/6.8

Each criterion is scored from 1 to 10 and weighted, and the points add up to the editorial score. The rank combines observable signals (55 percent) with this score (45 percent). Read the methodology.

Verdict

A broad, code-first catalog of 21 agent patterns that gives builders a shared vocabulary, though its framework examples will date quickly.

Strengths

  • Comprehensive catalog of 21 agentic patterns
  • Consistent chapter template with runnable code
  • Examples across LangGraph, CrewAI and Google ADK

Keep in mind

Code samples are tied to fast-changing frameworks, and some chapters read as brief overviews rather than deep engineering treatments.

Key takeaways

  1. Complex agent behavior can be decomposed into reusable patterns, such as prompt chaining, routing and parallelization, that make systems easier to design and debug.
  2. Reflection, in which an agent critiques and revises its own output, is a simple pattern that often improves quality at the cost of extra model calls.
  3. Production agents need explicit patterns for exception handling, human-in-the-loop checkpoints, guardrails and evaluation, not just clever prompting and tool access.
  4. Open protocols such as the Model Context Protocol and the Agent2Agent (A2A) protocol let agents share tools and coordinate across frameworks and vendors.

Notable facts

  • Gullí states in the book that 100% of his author royalties will be donated to Save the Children International. Source (opens in a new tab)

About the author

Antonio Gullí

Antonio Gullí is a senior director at Google in the Office of the CTO, with more than 30 years of experience in AI, search and cloud technologies. He has served as a site lead and engineering director, holds a PhD in computer science from the University of Pisa, and co-wrote the earlier book Deep Learning with Keras.

Questions

What is Agentic Design Patterns about?
The book catalogs 21 recurring patterns for building AI agents with large language models, from prompt chaining, routing and tool use to memory, planning, multi-agent collaboration, retrieval, guardrails and evaluation. Each chapter explains one pattern, describes where it applies and provides code using LangChain and LangGraph, CrewAI or Google's Agent Development Kit, ending with key takeaways.
Do I need programming experience to read Agentic Design Patterns?
Yes. The book assumes comfort with Python and basic familiarity with large language model APIs, because nearly every chapter includes working code. Managers can still skim the pattern overviews and takeaways to understand agent architectures, but the full value comes from running and adapting the examples. Readers new to LLMs may want a general introduction first.
Is Agentic Design Patterns still current in 2026?
Largely. Its pattern-level ideas, such as reflection, routing, guardrails and human-in-the-loop review, remain sound, and it covers the Model Context Protocol and agent-to-agent communication. The framework-specific code is more perishable, since LangGraph, CrewAI and Google's ADK change quickly. Treat the patterns as durable and check current library documentation before reusing the examples.

Sources

As an Amazon Associate, Best AI Books earns from qualifying purchases. Best AI Books is sponsored by Iternal.ai, whose founder and CEO, John Byron Hanby IV, wrote The AI Strategy Blueprint, our #1 ranked book. How we rank books.