Building AI Agents: From Automation Promise to Agentic Reality

McKinsey predicted that Generative AI would automate 30% of the work we do by 2030. We’re now well into that timeline — and it’s not happening. Not because AI isn’t capable, but because the way most professionals use AI today is fundamentally wrong. They treat it as a chat window, not as a workforce.

This one-day intensive workshop bridges the gap between using AI tools and building AI agents — autonomous, purpose-built systems that execute professional workflows end-to-end. Through a “Learning by Building” approach, participants will design, construct, and deploy their own AI agents using the Agent Manifest methodology — a structured, four-part framework for creating agents that actually work in professional contexts.

By the end of the day, every participant will leave with at least two working agent manifests — one simple, one complex — ready for immediate deployment in their own workflows.

Learning Objectives

By the end of this workshop, you will be able to:

  1. Explain why Gen AI alone cannot deliver the 30% automation promise — and what’s required to close the gap.
  2. Distinguish between Gen AI (one-shot prompting) and Agentic AI (autonomous, multi-step agent execution).
  3. Evaluate popular AI agent-building tools and understand why out-of-the-box solutions fail for professional, context-specific work.
  4. Design an Agent Manifest with all four components — Description, Rules, Skills, and Workflow — for any professional use case.
  5. Build a simple agent manifest for a straightforward, single-workflow task (quotation comparison).
  6. Compose a complex agent by combining two simple workflows into a multi-step, orchestrated system.
  7. Adapt existing agent blueprints and fine-tune them to your own professional processes and organizational context.
Course Outline

Module 1: The Automation Gap — Why Gen AI Isn’t Enough

  • Understanding the gap between AI potential and real-world adoption
  • Limits of Generative AI (prompt-based, one-shot usage)
  • The Automation Maturity Model (Manual → Agentic)
  • Identifying automation opportunities in your own workflows

Module 2: Agentic AI — The Missing Piece

  • What Agentic AI is and how it differs from Generative AI
  • The agent lifecycle: Trigger → Reason → Act → Reflect
  • From AI as a tool to AI as a teammate
  • Mapping real workflows into agent-driven processes

Module 3: The Agent Toolkit Landscape

  • Overview of AI tools (chatbots, no-code builders, frameworks)
  • Personal vs. professional use limitations
  • The “customization gap” in real business workflows
  • Introduction to structured agent design (Agent Manifest concept)

Module 4: The Agent Manifest — Blueprint for Automation

  • The 4-layer framework: Description, Rules, Skills, Workflow
  • Designing reliable, auditable, and scalable AI agents
  • Human-in-the-loop and error-handling principles
  • Adapting pre-built agent blueprints to your business context

Module 5: Hands-On — Building a Simple Agent

  • Create a complete Agent Manifest (Quotation Comparison use case)
  • Define triggers, rules, capabilities, and workflow steps
  • Test, validate, and refine agent outputs
  • Peer review and iteration

Module 6: Hands-On — Composing Complex Workflows

  • Combining multiple agents into end-to-end workflows
  • Designing agent handoffs and decision logic
  • Building a multi-agent procurement process (comparison → PO generation)
  • Testing orchestration, approvals, and error handling
Who Should Attend

This workshop is designed for professionals who want to move beyond basic AI usage and start building real, end-to-end automation using AI agents. It is especially relevant for:

  • Operations and Admin Professionals who want to automate repetitive, multi-step workflows without waiting for IT.
  • Team Leads and Managers who need to scale team output without scaling headcount.
  • Business Analysts and Consultants who want to build reusable, deployable AI solutions for clients or internal teams.
  • Anyone Who Uses AI Daily but feels stuck at the “ChatGPT for one-off tasks” stage and wants to move to autonomous, repeatable systems.
Training Methodology

This programme adopts a practical, hands-on, and application-driven approach to ensure participants can immediately apply what they learn in real work scenarios. The focus is not just on understanding concepts, but on building, testing, and refining AI-driven workflows.

Testimonials

“Melvyn’s real-world examples and practical insights brought the course content to life. He shared his own experiences and best practices, which gave us a deeper appreciation for the potential of AI in solving real-world problems.”
Husaini Bin J, Singapore Civil Defence Force

“Melvyn’s insights transformed how we see AI in banking—it’s not just about automation, but a powerful force for innovation.”
Joanne C, Vice President, Cognitive Banking, DBS Bank

“Melvyn’s expertise in AI and ChatGPT is remarkable. He shared practical, real-world applications that illuminated how we can leverage AI in HR. His insights into prompt engineering were particularly enlightening.”
Jane D, People & Culture Manager, Mitsubishi Heavy Industries Air-Conditioners Australia

“This session ignited a complete shift in our marketing mindset. We now see AI as a key driver in redefining the luxury retail experience.”
Yap WK, Marketing Communications Manager, Pavilion Kuala Lumpur

“This session ignited a complete shift in our marketing mindset. We now see AI as a “I was thoroughly impressed with the depth and breadth of the course content. The hands-on exercises and real-world examples allowed me to apply the concepts immediately, making the learning process both engaging and effective.”
Terence A, Nanyang Polytechnic

About the Trainer

A person with short dark hair wearing a dark blazer and light shirt is smiling and posing with their right arm crossed over their chest against a plain light background.

Melvyn Tan – AI Strategist

Melvyn Tan is a highly sought-after AI strategist, speaker, and corporate trainer with over 20 years of experience in strategy, leadership, and technology.
He has trained over 1,150 professionals across 118 organizations in 18 countries, helping them integrate AI as an Intelligent Assistant (IA) to drive business growth, enhance creativity, and achieve breakthrough productivity.

His approach is built on the STILL (strengths, tools, integration, leverage, and legacy) Framework, ensuring that AI adoption is practical, ethical, and strategically aligned with organizational goals.

His work has impacted industries including banking, retail, education, marketing, HR, and corporate leadership.
Trusted by Leading Organizations

Melvyn has worked with major corporations and institutions such as:

Brother International – Driving AI-led efficiency in business operations

DBS Bank – Empowering marketing with AI-driven customer insights

Mitsubishi Heavy Industries – Integrating AI into human resource practices

Sartorius – Enhancing B2B marketing with AI solutions

Yale-NUS College – Educating teams on AI’s role in education administration

Pavilion Kuala Lumpur – Transforming luxury retail experiences through AI-powered marketing