AMIA#31 | Generate contextual retrospective ideas with ChatGPT or Bing Chat
Aha! Moments in Agility
Live in your inbox every Wednesday to help you build valuable software sooner
AMIA: #31
Reading Time: ~ 3 minutes
Today, I’m going to show you how to use ChatGPT or Bing Chat to generate contextual, relevant and unique retrospective ideas.
This is an example that I used at work myself.
Initial prompt
I am going to run a retrospective for a group at work. I want you to list ideas for columns where the participants can put their thoughts. For example: What went well, What can improve, Next Actions. Generate 5 sets of 3 column headings for my retro.
Response
I wasn't happy with the response. It borrowed from traditional retro styles like rose, bud, thorn or start, stop, continue.
So, I gave it a follow up, more specific prompt.
Second prompt
Generate 5 more. This time don't borrow from traditional retrospective examples. I want to focus on highlighting information, fears and hesitations in the group towards an upcoming deadline. I don't have more information. Come up with ideas for the retro on your own.
The reason I added 'I don't have more information' is because it expected me to give permission to the web page. I didn't have a specific page to look at.
Response
Much better and contextually relevant response.
I see. You want some more creative and original ideas for your retrospective columns that can help you address the challenges and concerns of your group. Here are five more sets of three column headings that I came up with:
Know - Don’t Know - Need to Know: What do we know, don’t know, and need to know about the upcoming deadline and the project requirements?
Confident - Anxious - Curious: How do we feel about the upcoming deadline and the project deliverables? What are we confident, anxious, or curious about?
Strengths - Weaknesses - Opportunities: What are our strengths, weaknesses, and opportunities as a group in relation to the upcoming deadline and the project goals?
Expectations - Reality - Actions: What are our expectations, the current reality, and the actions we need to take to meet the upcoming deadline and the project standards?
Challenges - Solutions - Feedback: What are the challenges we faced, the solutions we implemented, and the feedback we received or need in the previous sprint and for the upcoming deadline?
I ended up using Expectations, Reality, Actions format which I know would have resonated with the group based on the conversations I'd heard.
Conclusion
Here are some tips you can use to modify or re-use for your purposes:
Have a conversational style when you chat with the large language model.
Give it a context, what you have want as an output, and use an example.
I usually add ‘before you generate, ask me any clarifying questions. If you understand what I need, say “Ready”’. This makes sure that the LLM is working step by step and can ask questions before going ahead.
The initial response is just that. If the direction is good, this is the time to iteratively start getting specific.
When you’re ready…
I can help you in 3 ways:
Check out the resources list
Forecast timelines like a Pro in under 1 minute - Zero Estimate System
Fix the root cause of bugs & defects in your team - Root Cause Analysis Template
Book One on One coaching with me - Become a go-to Scrum Master in your organization.
Find me on LinkedIn and introduce yourself :)