Early in 2025, I shifted how I use AI. Instead of browser-based tools, I moved to AI-integrated editors like Cursor and Windsurf. That shift made me step back and ask: Where can AI help in the Scrum framework?
Since then, I’ve been experimenting sprint after sprint. Each cycle, I try something new, learn from it, consolidate what works, and build on it.
Learning from Scrum.org’s AI Essentials
I recently took two classes from Improving: Professional Product Owner – AI Essentials and Professional Scrum Master – AI Essentials. Both were excellent—good balance of foundational knowledge and hands-on practice.
I want to understand how Scrum.org integrated AI into these accountabilities and how my experiments align with their approach. I also wanted to learn perspectives I hadn’t tried yet.
What Overlapped
Many things I’d already been doing showed up in the classes:
From Product Owner work:
-
Creating user stories from problem statements and conversation transcripts
-
Building personas
-
Refining stories with stakeholder feedback
From Scrum Master work:
-
Recording and transcribing all Scrum events (planning, daily scrum, review, retrospective)
-
Analyzing transcripts to pull out stories
-
Using AI to prepare for sprint planning, reviews, retrospectives, and backlog refinement
It was validating to see these practices shared in the classes.
What Was New
The Scrum Master class introduced me to Edward de Bono’s Six Thinking Hats—a structured thinking method where each hat represents a different mode: factual, emotional, cautious, and so on.
What resonated most was the idea of switching hats during a session. Start with one thinking mode, accomplish that intent, then switch to another. I’m adding this to my toolbox.
Shopify’s Mandate
I learned that Shopify has made the use of AI mandatory. Teams must prove they’ve used AI before requesting additional headcount.
My experience went the opposite direction. Our team was right-sized (we lost members), and we used AI to fill the gaps. We’ve done quite a bit successfully.
Setting Clear Expectations
The classes reminded me how important it is to define expectations—whether with AI or humans. If you ask a team member for something without clear expectations, you won’t get the results you want. Same with AI.
This is a skill worth honing in both contexts.
Facilitation and Workshops
One exercise covered creating a workshop to improve the team’s adherence to Scrum. This resonated with something I did late last year: preparing and facilitating a multi-week user acceptance testing period.
I used AI to prepare and facilitate UAT sessions, gather and process feedback, and improve the next session based on lessons learned. The pattern was the same: prepare, facilitate, analyze, improve.
Creating Videos
There was an activity to create a short video explaining Scrum events—like a one-minute video on sprint planning to get everyone aligned.
I tried using Gemini, but it capped videos at eight seconds (the activity called for a minute). Still, I was impressed by how quickly I created something that was heading in the right direction. It did give the person three hands, though.
Check out the video (and look for the 3rd hand)…
This sparked me to revisit my recent blog post about creating videos from stories. The tools we have now to streamline short, focused videos are worth exploring further.
What Humans Do That AI Cannot
We had good conversations about what humans can do that AI cannot, and vice versa.
One thing that came up resonates with something I’ve been telling others: reading the “unspoken” room. Humans can pick up body language, facial expressions, and patterns. We might notice that when Joe is quiet and starts blinking, he’s concerned but won’t speak up unless invited because he’s introverted.
AI can’t do that. But we can.
Human in the Loop vs. AI in the Loop
Some attendees were hearing about “human in the loop” for the first time—the importance of keeping humans involved in the AI process.
I shared my stance: I think less about “human in the loop” and more about “AI in the loop.” The Scrum framework is a human loop. The question is how AI fits into that.
It’s a different way to frame it, and I think it matters.
What I’m Taking Forward
These classes confirmed some of what I’ve been doing and introduced new tools, such as the Six Thinking Hats. They reminded me to keep setting clear expectations and to keep experimenting.
Sprint after sprint, I’ll keep asking: where can AI help us here?
Leave a Reply