Learn With Nathan Sign in
AI Agents book cover

Free Learn With Nathan Academy book

AI Agents

Understand, design, secure, monitor, and measure AI agents that can take useful action.

The learner understands chatbots but cannot design an agent that uses tools, follows boundaries, and completes work reliably. From AI conversations to supervised AI workers that pursue goals, use tools, and report their work.

The complete book is free. No fee or credit card required.

Author
Nathan Thiyagarajah
Chapters
20
Reading time
About 11 hours
Read a sample chapter

Book contents

Register to read all chapters for free.

Full sample chapters are available below without an account. Register free to read every chapter in this book and access the complete Academy. No fee. No credit card required.

  1. 01
    What Is an AI Agent?Explain an AI agent as a model inside a controlled loop with instructions, tools, environment, state, limits, and oversight.
    Read sample
  2. 02
    AI Agent vs Automation: What's the Difference?Choose fixed automation, AI-assisted workflow, or agentic execution based on path variability, consequence, evidence, and operating need.
    Read sample
  3. 03
    Build Your First AI AgentBuild a first agent around one bounded job, read-only tools, explicit completion, representative evaluations, and full trajectory review.
    Read free in Academy
  4. 04
    20 AI Agents Every Business Should ConsiderUse a list of 20 agent ideas as a portfolio of hypotheses, then rank only the few with real value, readiness, ownership, and control.
    Read free in Academy
  5. 05
    Your First AI EmployeeFrame a recurring agent as a supervised digital worker with a role charter, service limits, permissions, workload, review, and accountability.
    Read free in Academy
  6. 06
    Build an AI Marketing AgentBuild a marketing agent that coordinates evidence-backed campaign work while people retain claims, rights, audience, spend, publication, and brand authority.
    Read free in Academy
  7. 07
    Build an AI Sales AgentBuild a sales agent around verified account state, buyer value, consent, approved claims, commercial limits, and seller ownership.
    Read free in Academy
  8. 08
    AI Customer Service AgentsBuild a service agent that resolves only defined cases, verifies identity, follows policy, protects data, and hands off uncertainty with context.
    Read free in Academy
  9. 09
    AI Research AgentsBuild a research agent that plans dynamically while preserving source quality, claim support, disagreement, budgets, and honest incompleteness.
    Read free in Academy
  10. 10
    AI Operations AgentsBuild an operations agent around verified states, safe work orders, bounded tools, dependencies, approval, and recovery.
    Read free in Academy
  11. 11
    AI Voice AgentsDesign a voice agent with clear identity, consent, accessible turn-taking, narrow intents, confirmation, transfer, recording, and synthetic-voice controls.
    Read free in Academy
  12. 12
    AI Agents With MemoryGive an agent only the memory needed for continuity, with source, scope, access, conflict, correction, review, expiry, and deletion.
    Read free in Academy
  13. 13
    Giving AI Agents ToolsGive agents small, well-described tools with narrow permissions, validation, confirmation, idempotency, logs, quotas, and revocation.
    Read free in Academy
  14. 14
    Human-in-the-Loop AgentsDesign human control around material uncertainty and consequence with useful evidence, real options, ownership, and manageable review volume.
    Read free in Academy
  15. 15
    Multi-Agent SystemsUse multiple agents only when specialist separation improves quality or control, with explicit contracts, shared state, verification, and orchestration limits.
    Read free in Academy
  16. 16
    Agent MonitoringMonitor agent outcomes and trajectories for quality, policy, tool use, cost, latency, drift, exceptions, and incidents—not just uptime.
    Read free in Academy
  17. 17
    Agent SecuritySecure an agent across goal, instructions, data, identity, tools, memory, environment, supply chain, logs, and human recovery.
    Read free in Academy
  18. 18
    Measuring Agent ROIMeasure agent value per accepted outcome after autonomy, review, tools, failures, latency, security, change, and adoption are counted.
    Read free in Academy
  19. 19
    Agent-to-Agent CommunicationDesign agent handoffs with authenticated identity, structured messages, provenance, scoped authority, validation, timeout, error, and human escalation.
    Read free in Academy
  20. 20
    From AI Agents to Agentic SoftwareMove from one agent demo to agentic software through stable capability contracts, evaluation, security, observability, change control, and retirement.
    Read free in Academy