Module 14: Agent Builder
Lesson 14.3: Building Your First Agent
Lesson Promise
Turn a focused idea into a practical first Agent Builder prototype.
Real-World Scenario
A customer success team wants a helper that answers renewal-process questions from one approved playbook and a short FAQ.
Core Concept
A first agent should be intentionally small. A narrow agent teaches better design faster than a sprawling one.
The build path typically starts from the Microsoft 365 Copilot app or Agent Builder entry point where enabled, then adds description, instructions, starters, and knowledge.
The prototype is not complete until it has been tested against real questions and reviewed by the content owner.
Step-By-Step Workflow
- Open the agent creation experience where enabled in your Microsoft 365 environment.
- Name the agent around the job, not a cute internal nickname.
- Add a description that tells users exactly what the agent can and cannot help with.
- Add instructions for answer structure, tone, source use, uncertainty, and escalation.
- Add approved knowledge sources.
- Create conversation starters and run the first test set.
Prompt Lab
Bad Prompt
Make the agent helpful and friendly.
Better Prompt
Answer renewal-process questions using the renewal playbook and FAQ. If the answer is not in those sources, say so and route to the account owner.
Expert Prompt
Create the full setup brief for a renewal-process Agent Builder prototype: name, description, audience, knowledge, detailed instructions, answer format, conversation starters, refusal rules, escalation path, first ten test prompts, and launch risks.
Hands-On Exercise
Draft the setup brief for a small agent you could safely prototype.
Deliverable
A first-agent setup brief ready to paste into the builder.
Agent Builder Review Checklist
Common Mistakes
- Building a broad general helper instead of a focused role-based agent.
- Adding messy or outdated files and expecting the agent to infer what matters.
- Skipping refusal instructions for policy, legal, financial, HR, or customer-impacting questions.
- Testing only the happy path and missing vague, hostile, or sensitive prompts.
- Sharing an agent before ownership, feedback, and update routines are clear.
Quiz / Checkpoint
Why should the first agent be small?
Small agents are easier to source, test, govern, explain, and improve.
Official Sources To Verify
Want your AI product explained to a large AI-native audience?
Kingy AI helps AI companies turn complex products into clear, useful YouTube videos that drive awareness, product understanding, demos, clicks, and search visibility.
