Production AI Agents

6 Interactive Sessions

Short, interactive sessions: watch it work, try it, then build it yourself. You can go deeper at any time in the written chapters.

1. 1

   Tools: how an AI agent takes action

   A model can write a good email to Acme's VP, but it cannot send it, open the CRM or check a calendar until you give it tools. A tool is a written contract the model reads before it acts.

2. 2

   The loop: how an agent plans, acts and checks

   One model call cannot do the Acme task. The agent has to plan the steps, run a tool, look at the result and decide what to do next. That loop, written as a graph you can inspect, is called orchestration.

3. 3

   MCP: build your tools once, use them in any AI app

   You built six good tools for your agent. Now people want them in Claude Desktop, in their code editor and in a dashboard. MCP lets you offer them once, through one standard, and every app can find and call them.

4. 4

   Guardrails: a person approves what cannot be undone

   The agent is about to email Acme's VP and update the CRM on its own. Wrong contact? A made-up price? A sent email cannot be unsent. So the agent proposes, and a person approves anything that cannot be undone.

5. 5

   The agent app: show every step before it acts

   An agent that says “I did 12 things” is a black box nobody can trust. A good one shows each tool it wants to call, with the exact details, and asks before it acts. The app is where the approval gate becomes something a person can see.

6. 6

   Running the agent: keep it honest after launch

   The agent is live and acting for the whole sales team. How would you know if it started getting worse because a tool changed or a prompt drifted? Watch it, test it, roll it out in stages, and learn from every failure.

How this certificate is scored

Your certificate shows how many of your written answers met the rubric below, checked by an AI grader. Anyone who checks your certificate sees it.

• You answer 6 short questions, one for each module, in your own words (at least 40 words each).
• An AI grader checks each answer against what its module asks for: it must answer the question, use the course's practice case specifically, and be correct.
• To pass, at least 5 of the 6 answers must meet that bar. The certificate says the answers were graded by AI.

The rubric

1. Module 1, Tool Design: 6 tools with JSON Schema validation.
2. Module 2, Agent Orchestration: LangGraph plan→execute→observe.
3. Module 3, MCP Server: Universal tool adapter.
4. Module 4, Guardrails: Approval gates + audit trail.
5. Module 5, Agent App: Chat UI with tool visualization.
6. Module 6, Production Pipeline: Monitoring + staged rollout.

What you walk away with