AI Agent Development: Build Your First AI Agent
Three hours, live online, on September 28, 2026. Work through one shared business scenario using Claude chat, Cowork, and Claude Code, then plan how to apply what you learn to your own work.
For the September 28 session, a paid Claude plan and advance setup are required. Contact me for preparation instructions before the workshop. If your organization restricts software installation, contact me to arrange a browser-based participation path.
Why this workshop
An agent needs a clear task, useful knowledge, instructions, and boundaries. You will use the same business scenario across three ways of working with Claude, compare the results, and see how the setup changes what the AI can do.
I have run this workshop twice, in June 2026. What surprised me most was where the difficulty sits. People understand what an agent is quickly. The hard part is deciding which task in your own work is worth handing over, and what you do when the agent gets it wrong. The September session builds on that experience with a shared scenario and time to apply the lessons to your own work.
What you will learn
- Understand the AI agent landscape and where agents create value.
- Tell chatbots, agents, and agentic workflows apart.
- Design an AI agent on a structured eight-field design canvas.
- Evaluate and communicate business impact and return on investment.
- Apply risk mitigation and governance principles.
- Navigate the current AI tools with confidence.
How the three hours run
Understand the agent
See how a model, instructions, knowledge, and tools work together.
Start in chat
Work through a fictional business inquiry using Claude chat and a shared set of materials.
Work with Cowork
Use the same scenario in Cowork and compare the workflow and output.
Build with Claude Code
Carry the scenario into Claude Code and examine the role of persistent instructions, knowledge, and guardrails.
Test the boundaries
Use a second inquiry to test how the agent handles constraints and requests outside its remit.
Apply it to your work
Use the AI Agent Design Canvas to plan an agent for a recurring task in your own work.
What you leave with
Who it is for
Leaders in strategy, operations, innovation, and digital transformation. C-suite executives and mid-senior leaders exploring AI adoption. Business unit leaders who want to know what AI agents can do for their teams. No technical background is required. The workshop is designed for business leaders, not engineers.
Practical details
- September 28, 2026
- 5 to 8 PM Eastern
- Live online (Zoom)
- 3 hours
- Paid Claude plan and advance setup required; contact the instructor for instructions.
- 20 to 25 participants
- $500
- Certificate of completion, University of Miami
- September 20, 2026
What participants said in June
"Excellent and useful program! The three hours flew by."
"Very useful overview and ways to think about tasks and improving AI support for specific tasks and specific duties."
Enroll or ask
Enrollment and payment run through University of Miami Business School Executive Education. For questions about enrollment, write to Execed@miami.edu. For questions about the content, or whether the session fits you or your team, write to me directly.