THE AGENTIC WORKSHOP: Orchestrate Specialised AI Agents from First Idea to Finished Outcome

The Agentic Workshop is a first-principles guide for non-technical readers who want to understand AI agents without being buried in code, jargon or exaggerated promises. It begins with a human need, then opens the machinery one layer at a time: goals, roles, context, observation, decisions, tools, feedback, memory, permissions, budgets, handoffs and stopping conditions.

Through clear explanations, natural workshop dialogues and practical exercises, Ravindra Nayak shows how a vague idea can become a bounded mission; how one agent can observe, decide and act within limits; and how several specialised agents can research, plan, build, test, review and communicate around one shared outcome. The book also explores the harder questions that are often left out of demonstrations: What happens when agents disagree? Which information deserves to be remembered? How should confidence be measured? What must be logged? Who approves consequential actions? How should a system recover, escalate or stop?

Readers move from simple foundations to dynamic planning, multi-agent orchestration, selective memory, evaluation, least-privilege permissions, audit trails, incident response and long-term stewardship. Practical mathematics is introduced gently through rates, thresholds, priorities, costs and evidence—never as a barrier, but as a way to think more clearly.

The final Value Edition turns the entire book into a mental training workshop. It helps readers frame problems, break complexity into meaningful chunks, ask higher-value questions, recognise patterns, reason with constraints, test counterexamples and design a complete mission from idea to responsible outcome.

This is not a book about removing people from meaningful decisions. It is about placing human judgment where it matters most while allowing intelligent systems to support the work they can perform safely and visibly.

By the end, readers will be able to describe an agentic mission in plain language, assign specialised responsibilities, design handoffs, choose appropriate levels of autonomy, evaluate results, govern memory, create approval gates and build clear rules for completion, recovery and safe stopping.

For creators, educators, professionals, team leaders and curious beginners, The Agentic Workshop offers a calm path into agentic AI: understand the mission, design the system, supervise the work and remain responsible for the outcome.