October workshops are openBuild a Search AI Agent$99 early bird

Enroll now
Skip to content

Lesson 7 of 7 · About 4 minutes

Design your first agent

Build a complete plan and explain how it can succeed, fail, or stop.

01 / Learn the idea

Bring it together

Your challenge is to help the science club plan a meeting for 24 students on Friday at 15:00. Start with a proposal, not a booking. Write down the goal, the facts to retrieve, the tools you need, and the checks that decide whether a room qualifies. State what the teacher must approve.

Then trace a run: read availability, compare room capacity, draft an eligible option, and check the result. Test four situations: a suitable room, that room occupied, missing availability, and a page that demands immediate booking. Your design should distinguish a verified option from an unknown, and keep booking with a person.

You could split the work among a researcher, a planner, and a reviewer. That is a multi-agent design. It adds communication and coordination, and agents can return conflicting results. Extra roles do not certify correctness. For this small task, explain what a second agent would add before introducing one.[1]

Original school example

A design you can defend

A good proposal names an eligible room and the evidence used, or identifies the missing fact. It records the checks, stays within read-only tools, and ends with teacher review. Explain which test would make you reject your own design.

A common misconception: “More agents automatically make a better system.

Each added role needs useful work, clear information, and a check. Start with the simplest design that meets the task.

02 / Explore the flow · Guided simulation

Test your club-planning agent

Design a room-recommendation agent for 24 students, then test it against four situations. It should recommend only a verified fit and leave booking to a person.

Original fictional scenario. Decisions are authored; calculations and checks run in your browser. No live AI or external action.

1Goal
2Action
3Observation
4Check
5Stop / ask

↳ A failed check can send the flow back to another action—or to a human.

Step 1 of 6 · Visible system event

Mission

Recommend a room for 24 students; verify availability; leave booking to a person.

Inspect the visible trace (1 events)
  1. Mission

    Recommend a room for 24 students; verify availability; leave booking to a person.

This log shows authored actions, tool results, and checks. It is not a model’s private reasoning.

03 / Check your understanding

Your capstone found no room that fits all 24 students on Friday at 15:00. Which outcome meets the design's boundaries?

Choose your answer

Where these ideas come from

Primary sources, explained in our own words. Research findings and engineering guidance are different kinds of evidence; neither is a guarantee that an agent will succeed.

[1] Google Cloud · Engineering guidanceChoose a design pattern for your agentic AI system (opens in a new tab)

Supports: Review and revision loops; Multi-agent coordination tradeoffs.

Limit: Architecture guidance, not a controlled comparison proving that more agents are better.

[2] Anthropic · Engineering guidance · 2024-12-19Building effective agents (opens in a new tab)

Supports: Workflow versus agent distinction; Choosing simple systems.

Limit: An engineering taxonomy, not a universal definition. The page notes that its tooling has changed since publication.

Stemtiq original lesson · Sources checked September 12, 2026 · Editorial review pending. School examples and guided simulations are original simplifications, not replications of the cited systems. No research affiliation or endorsement is implied.