Skip to content
Page 9 of 10

Reflection 1

How This Reflection Works

Your Facilitator will pull from these questions to lead a cohort-wide discussion. If your AI assistant is building during a Challenge and you have a free moment, these also make great team conversation starters while you wait.

What You Built

  • Walk us through what you built. What's the piece you're most proud of? What made it work?
  • When you give AI a well-scoped prompt with clear intent and structure, what do you expect it to get right on the first try? What do you expect it to get wrong? If your team built with AI during the challenge, did the reality match those expectations?
  • When working with AI, where do you think teams typically spend the most time: getting a first version out, or refining it into something they're happy with? What does that tell you about how effort is distributed when AI is doing the drafting?

What You Practiced

  • Think about the Three Pillars: Scope, Intent, and Structure. Which of these do you think most people neglect when they first start prompting? Can you think of an example (from the challenge, from your own experience, or hypothetically) where being more specific on one pillar would make a noticeable difference in the output?
  • Lesson 1 introduced user stories with acceptance criteria as a way to communicate with AI. Why might structured requirements produce different results than plain English descriptions? What do acceptance criteria give AI that a casual request doesn't?
  • The Explore, Plan, Implement, Verify cycle includes a deliberate verification step. Why is that step easy to skip? What kinds of problems does verification catch that a quick glance at the output would miss?
  • AI gives you different answers each time, even for the same prompt. When is that variation an advantage, and when does it become a problem? How does being more specific in your prompts change how much variation you see?

How You Worked

  • How did your team organize? Did everyone gather around one screen, or did you try different approaches? What would you do differently next time?
  • Was there a moment where a teammate's suggestion changed the direction of what you were building?

Looking Ahead

  • When you build something inside a chat tool, the output lives only in that conversation. What are the biggest limitations of that approach, and what's the first thing you'd want to change to move from a prototype to something real?