About the Course

AI Agent Manager is a practical, role-based course on designing, managing, and evolving AI-assisted knowledge work through the lens of flow, constraints, and system thinking.

This course is built on a simple idea: AI does not fix broken systems. It amplifies them. Most AI failures are not caused by poor models, but by poor workflow design, weak architecture, and lack of management discipline.

Across four structured modules, you will learn how to identify where AI belongs, how to apply it safely and effectively, and how to evolve AI-enabled systems over time. Instead of focusing on tools, this course equips you with a durable approach to making AI work in real environments.

Who Should Attend?

This course is designed for professionals responsible for outcomes, workflow performance, and AI adoption in real systems.

It is especially valuable for:

• Product and engineering leaders
• Developers working on AI-enabled solutions
• Managers responsible for operational efficiency
• Agile coaches and transformation leaders
• Architects designing AI-supported workflows

If you are accountable for making AI deliver measurable value, this course is for you.

Key Objectives

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

• Understand why AI initiatives fail and recognize that AI value is fundamentally a workflow problem
• Map workflows, uncover hidden work, and distinguish between functions and flow
• Identify bottlenecks and constraints, and understand their impact on system performance
• Apply constraint-first thinking by addressing bottlenecks before introducing AI
• Use root cause analysis and cheapest-first interventions to improve flow
• Select the right technology for the problem, recognizing that LLMs are not always the default
• Compare and apply rules, retrieval, machine learning, and automation appropriately
• Design AI architectures including workflows, agents, memory, orchestration, and control points
• Design for stability before scaling AI systems
• Implement security controls such as filters, permission boundaries, sandboxing, and human approvals
• Improve reliability using WIP limits, structured outputs, decomposition, and evidence grounding
• Define and apply explicit policies, including templates, validation, and enforcement mechanisms
• Govern and evolve AI systems over time
• Run safe-to-try experiments with versioning, rollout, and rollback strategies
• Measure performance using flow, quality, risk, and outcome metrics
• Design effective feedback loops at micro, workflow, and system levels
• Establish governance models with clear roles and alignment to standards such as ISO/IEC 42001

What Will You Do Differently After?

After completing this course, you will shift from a tool-first mindset to a system-first approach.

You will:

• Diagnose workflows before introducing AI
• Focus on constraints and flow as the primary drivers of improvement
• Make deliberate decisions about when AI should and should not be used
• Design AI-enabled workflows that are stable, secure, and reliable
• Apply policies and controls that make AI behavior predictable and manageable
• Continuously improve AI systems through experimentation, metrics, and feedback
• Establish governance practices that support long-term, scalable AI adoption

Certification

Participants who complete this course will receive certification from Kanban University, validating their capability to design, manage, and evolve AI-assisted systems in the role of AI Agent Manager.