Reflection 1

How This Reflection Works

Below is a question bank. As a team, look through the questions and pick two or three that connect to what you just experienced during the build. Discuss them among yourselves. There are no wrong choices; pick the ones that spark the most interesting conversation for your group.

1. What surprised you most about building with AI? What matched your expectations, and what didn't?

2. You wrote user stories with acceptance criteria before building. How did that change the output compared to giving AI a plain English description? Were there moments where the acceptance criteria caught something you would have missed otherwise?

3. AI is probabilistic: it predicts the most likely response rather than looking up a correct answer. Did you notice moments where AI was confidently wrong? How did you catch it?

4. If you set up a context file, how did it change the quality or consistency of what AI produced? If you didn't get to one, what context did you find yourself re-explaining across conversations?

5. AI produces different output each time, even for the same prompt. When was that variation helpful, and when did it get in the way?

6. Where did your team spend most of its time: getting a first version out, or refining it into something you were satisfied with? What does that tell you about how effort is distributed when AI does the drafting?

7. Think about how your team organized during the build. Did everyone gather around one screen, or did you split up the work? What would you change for the next sprint?

8. If someone in your organization who has never used AI for building software asked you what it was like, what would you tell them?