← Blog

Leaders With AI Blog

Scrum Master Role in Mentoring Teams for AI Fluency

Published June 11, 2026 / By Joshua Partogi
scrum master mentoring stance artificial intelligence

The Art of Wearing Different Hats


A skilled Scrum Master is not a one-dimensional facilitator. They are a shapeshifter, moving fluidly between the roles of teacher, coach, facilitator, and mentor depending on what their team needs in any given moment. In my experience, one of the most misunderstood and underused stances is the mentor hat.

People often conflate mentoring with coaching. I'll admit I use them interchangeably in casual conversation, but understanding the distinction matters, especially when your team is navigating something genuinely new, like integrating AI into their everyday work.

Mentoring vs. Coaching: Why the Difference Matters


The International Coaching Federation (ICF) draws a clear line: in pure coaching, the coach does not give advice. The reason is partly ethical. If someone acts on your advice and it leads somewhere harmful, that's a real problem. The coach's role is to ask powerful questions that help the coachee discover their own answers.

Mentoring is different. As a mentor, you bring your experience. You share your war stories, your hard-won lessons, your perspective on "what worked for me when I faced something like this." The mentee can take that advice, leave it, adapt it, or use it as a springboard to something entirely their own.

The key is non-attachment. Just like a coach, a good mentor doesn't grip their advice too tightly. You offer the perspective. You give the disclaimer: "This is what worked for me, but that was a decade ago, and your context is different." And then you let go.

That disclaimer, counterintuitively, is empowering. When someone knows you're not expecting them to follow your advice to the letter, they feel free to think rather than comply.

In practice, I use a blend. Pure coaching (the kind where I ask only questions and give zero input) happens maybe 5% of the time, usually when someone specifically expects a formal coaching engagement. Most of the time, I mix coaching questions with mentoring wisdom, with facilitation, with teaching moments. Knowing which blend to use is itself the craft.


Enter the 4D AI Fluency Framework


As AI tools become embedded in how teams work, Scrum Masters face a new challenge: helping their teams navigate AI well. Not just technically, but ethically, thoughtfully, and sustainably.

The 4D AI Fluency Framework provides a map for this journey. Each "D" represents a distinct dimension of AI literacy that teams need to develop:

1. Delegation: Know What to Hand Off (and What to Keep)

The first dimension is about discerning which tasks are genuinely suitable for AI and which still require human judgment. This sounds simple, but it isn't. Teams often default to one of two failure modes: delegating too much (trusting AI with decisions that deserve human deliberation) or too little (refusing to let AI handle repetitive work that wastes human capacity).

Developing good delegation instincts takes practice, reflection, and often a mentor who has made the mistakes already.

2. Description: Prompt Effectively

The second dimension is about communication with AI. Prompting is not a technical skill in the traditional sense; it's a language skill. A single word in a prompt can dramatically change the output. Understanding how to frame a request, provide context, use examples, and specify constraints is a craft that improves with iteration.

This is where the mentoring stance shines. A Scrum Master who has worked through dozens of prompting experiments can offer heuristics that a team would otherwise spend weeks discovering on their own.

3. Discernment: Read the Output Critically

The third dimension is arguably the most important: the ability to evaluate what AI produces with a critical eye. Hallucinations happen. Biases get encoded. Sometimes the problem isn't the AI; it's the prompt. A biased or poorly-framed prompt will produce biased or misleading output.

Discernment is not skepticism for its own sake. It's the disciplined habit of asking: Does this make sense? What assumptions are baked in here? What might be missing? This is a habit that experienced practitioners model, not just teach.

4. Diligence: Use AI Ethically and Responsibly

The fourth dimension is about integrity. AI fluency without ethical grounding is just sophisticated carelessness. Diligence means asking who is affected by AI-assisted decisions, whether the data used is appropriate, whether the team's use of AI aligns with organizational values and human dignity.

This dimension connects directly to what I call the "AI Ethical Guardian" stance, a responsibility that Scrum Masters are uniquely positioned to hold for their teams.


Why the Mentoring Stance Is Uniquely Valuable for the 4D Framework


Here is the core argument: the 4D framework is not primarily a technical framework. It's a judgment framework. And judgment is not taught through instruction alone. It's developed through experience, reflection, and guidance from someone who has been there.

Consider what each dimension actually demands:

  • Delegation requires contextual wisdom about when human oversight matters. No policy or checklist can fully encode this. It comes from having seen what happens when the line is drawn in the wrong place.
  • Description requires iterative intuition about language and context. A mentor can share patterns of what works and why, saving teams from reinventing the wheel.
  • Discernment requires a trained skeptical eye. This is precisely where a mentor's war stories are most valuable: "Here's a time I was fooled by confident-sounding AI output. Here's what I learned."
  • Diligence requires moral seriousness that is modeled, not mandated. A Scrum Master who personally practices ethical AI use and talks about it openly, including the hard cases, shapes team culture more than any training module.

In each case, what the team needs is not more information. They need perspective from someone who has wrestled with these questions in real situations and is willing to share, without requiring that the team simply copy their conclusions.


The Mentor Who Doesn't Cling to Their Own Advice


There's a paradox at the heart of good mentoring: the most effective advice-givers are the ones least attached to having their advice followed.

When a Scrum Master mentoring their team on AI fluency says, "Here's how I've approached this, but it might not be right for you, and things have changed since I first tried it", they do something powerful. They model intellectual humility. They demonstrate that even experienced practitioners remain learners. They signal that the team's own judgment matters.

This posture is particularly important in the AI space, where the landscape is shifting so rapidly that any fixed set of rules is guaranteed to be incomplete. What teams need is not a definitive AI playbook. They need the capacity to build their own judgment, scaffolded by a mentor who respects their autonomy enough to trust them with it.


Putting It Into Practice


If you're a Scrum Master looking to apply this in your team context, here are a few starting points:

On Delegation: Share a story about a time you over-delegated or under-delegated to a tool (AI or otherwise). What did you learn? Invite the team to reflect on their own experiences before establishing any norms.

On Description: Experiment together. Run the same prompt with slight variations and compare outputs. Let the team discover the craft for themselves, and share your own hard-won heuristics as inspiration, not instruction.

On Discernment: Make it safe to question AI output. When the team reviews AI-assisted work, model the habit of asking "what assumptions does this output make?" Share examples where you were initially fooled.

On Diligence: Bring ethical questions into retrospectives and planning conversations. "Are we comfortable with how we used AI this sprint? Is there anything we'd do differently?" This isn't policing; it's cultivation.


Conclusion


The 4D AI Fluency Framework (Delegation, Description, Discernment, Diligence) gives teams a vocabulary for developing genuine AI fluency, not just AI familiarity. But a vocabulary without lived practice is just jargon.

The Scrum Master wearing the mentor hat is the bridge between the framework and its real-world application. They bring experience. They share wisdom. They disclaim freely and let the team run with their own interpretations. And in doing so, they help their teams develop not just AI fluency, but the confidence to keep developing it long after any training ends.

That's what good mentoring looks like. Not handing over answers, but lending a perspective and trusting the team to do something better with it than you ever could.

Joshua Partogi is a Professional Scrum Trainer and facilitator of PSM AI Essentials. This post is adapted from session notes and expanded with commentary.

Related reading