When considering the use of Artificial Intelligence, we often jump straight to asking: What can AI do for me? That question is only a stone’s throw from a more unsettling one: Will AI make my role redundant?
For Scrum Masters and Agile Coaches, perhaps both questions start in the wrong place. Perhaps we should instead ask: What can AI take away, so that I can spend more time on the work that requires me to be human?
Generative AI (GenAI), as used by ChatGPT, Claude, Gemini, Copilot, etc., can draft, summarise, organise, analyse and suggest ideas extraordinarily quickly. Used well, it can make a Scrum Master more efficient and potentially improve the quality of their work. There are also an increasing number of AI tools (and even “AI Scrum Master” agents) promising to revolutionise Agile work.
But Scrum Mastery is fundamentally concerned with people, interactions, learning and change. Those are areas where delegating too much to AI is counterproductive.
The opportunity, therefore, isn’t to create an AI Scrum Master. It is to use AI selectively – accelerating some work and augmenting thinking, while deliberately keeping humans in the loop.
Think of these as three modes for the AI-enhanced Scrum Master: Accelerate, Augment and Stay Human.
The 2001 Agile Manifesto even has something to say on the subject!
Individuals and interactions over processes and tools
Accelerate: Spend Less Time on Scrum Master Mechanics
There are plenty of activities surrounding a Scrum Master’s work where performing the mechanics of the activity isn’t particularly valuable.
Imagine you’ve collected notes from a workshop and need to organise them into themes. Or you need to turn rough notes into a concise stakeholder update. Perhaps you’re designing a retrospective and want five alternatives to the format you’ve used recently.
These are good candidates for AI assistance.
AI might help you with:
- automating repetitive tasks;
- summarising or converting information into a different format, or for a different audience;
- drafting potential User Stories from interview notes;
- suggesting questions or possibilities to explore during Backlog Refinement;
- reviewing output from Sprint Planning, perhaps highlighting potential bottlenecks or over-commitment;
- checking Acceptance Criteria for consistency and completeness;
- summarising outcomes from a Daily Standup (consider AI tools for automated note taking);
- analysing team metrics, perhaps identifying trends in aspects such as velocity, or supporting delivery forecasting and delivery predictability);
- generating visualisations for Sprint Review reports;
- suggesting ideas for workshops or Retrospective formats;
- creating facilitation materials;
The important qualification to all of these is: AI produces the first draft. AI can dramatically reduce the time required to get from a blank page to something useful. The Scrum Master can then review, challenge and adapt the result using their knowledge of the team and organisational context.
Critically, the Scrum Master should not delegate their judgement, responsibility or accountability. Instead, the key is to reduce the mechanical effort involved in producing the output.
And the time saved matters. Twenty minutes saved writing a report is twenty minutes that could instead be spent talking to a team member, observing how people are working together, investigating an impediment, inspecting and adapting (this is Agile, after all!). The value isn’t simply in saving twenty minutes. It’s what the Scrum Master chooses to do with those twenty minutes.
Augment: Use AI to Aid Your Scrum Master Thinking
An interesting, but perhaps riskier, proposition is asking AI to help us think about our work. To lower the risks, remember that AI should be thinking with you, not for you.
Sprint Retrospectives – Example
Suppose participation in your Sprint Retrospectives has declined. You have a theory about why. Instead of asking AI to diagnose the team, you might say:
“I think participation in our Sprint Retrospectives is declining because the team doesn’t believe anything changes afterwards. Give me five alternative hypotheses and associated questions I could ask to investigate each one.”
AI doesn’t know why your team has become less engaged. But it can help prevent you becoming overly attached to your own explanation. Notice especially how you are not asking for an answer.
Conflict Resolution – Example
Similarly, before a difficult conversation, you could describe an appropriately anonymised situation and ask AI to challenge your interpretation, suggest questions you might ask, or role-play possible responses.
“What assumptions might I be making about this situation? Give me three alternative ways the people involved might perceive it.”
AI hasn’t suddenly developed empathy or an understanding of your colleagues. It is generating plausible alternatives. You determine whether any of them deserve further investigation.
An AI-Enhanced Scrum Master, not an AI Scrum Master
AI can generate possibilities quickly, challenge our initial thinking and help us approach a situation with greater curiosity.
Used this way, AI becomes less like an oracle and more like a sparring partner – or perhaps, for those with development experience, a slightly more vocal rubber duck!
Stay Human: Don’t Delegate What Matters
There is another category of Scrum Master work where efficiency shouldn’t be our primary objective: the human side.
Consider listening to somebody describe a problem. Technically, there may be faster ways to obtain the information. You could ask for a written summary and have AI analyse it. But obtaining information isn’t necessarily the only purpose of the conversation.
Listening builds relationships. Follow-up questions reveal things we hadn’t anticipated. Tone, hesitation and body language provide context. The person speaking may understand their own problem differently by talking it through. The apparent inefficiency of human conversation is sometimes where much of its value lies.
The same applies to coaching, navigating conflict, building trust and facilitating difficult conversations. AI might help a Scrum Master prepare for these interactions, but it shouldn’t become a convenient way of avoiding them.
There is a significant difference between asking AI, “What questions could help me prepare for this difficult conversation?” and asking it to determine what you should say to somebody based on its assessment of their personality.
AI should not be treated as knowing what people think.
Giving an AI a selection of comments from a team and asking it to identify possible themes may be useful. Asking it to determine whether “the team lacks psychological safety” is quite different. That is a conclusion requiring investigation, context and human judgement.
AI-generated hypotheses should remain hypotheses until human judgement has been applied.
Don’t Let AI Take Ownership Away from the Scrum Team
There is another risk particularly relevant to Scrum Masters: AI makes it extremely easy to provide answers.
Imagine a Scrum Team encounters a recurring impediment. Within seconds, a Scrum Master could describe it to an AI system and return with ten possible solutions.
That sounds helpful. But should they?
Perhaps the valuable activity is the team discussing the problem, exploring alternatives, disagreeing, experimenting and eventually deciding what to try.
The Scrum Master’s AI-generated list might contain excellent ideas while simultaneously taking away something more important: the team’s ownership of the problem and its solution.
The same problem could arise in a Retrospective. Feed the data into AI and it may rapidly produce themes, conclusions and recommended actions. But if the team could have discovered those things themselves through conversation, what have we gained – and what have we lost?
Sometimes the Scrum Master’s job isn’t to provide an answer, even when technology makes providing one almost effortless.
AI could instead be used afterwards to challenge the team’s thinking, identify an alternative they haven’t considered, or help develop an experiment around an idea the team has already generated.
The team still owns the problem, the thinking and the decision. AI is simply another resource it can draw upon when useful.
The difference is subtle but important: AI supports the team’s thinking rather than replacing it.
This is perhaps a better way to think about keeping a human in the loop. Rather than putting the Scrum Team in the loop of an AI process, put AI in the loop of the Scrum Team’s process.
Without that perspective, the risk to both ownership and learning is very real. Studies are already available that indicate the negative effect of overuse of AI on learning and knowledge retention.
Always Keep a Human in the Loop of AI
Aside from the perspective that only humans can bring, generative AI has other limitations that make human oversight essential.
It can produce incorrect information convincingly (known as AI hallucination), reproduce biases contained within its training data or introduced through the way we frame a question, and present polished answers that create a false sense of authority.
There are also important confidentiality considerations. Team conversations may involve commercially sensitive information, personal data or observations about individuals. Scrum Masters should understand their organisation’s policies and the data protections provided by any AI system before sharing such information.
Anonymisation can help, but the broader principle is simple: being able to paste something into an AI system doesn’t mean you should.
We should also be alert to automation bias – our tendency to give additional credibility to a conclusion simply because a system generated it.
AI can suggest, but Humans must judge. Humans remain accountable.
Could It? Should It?
So how does a Scrum Master decide when and how to involve AI? Before using it for an activity, ask four questions:
Could AI do this?
Is this something AI can genuinely help with?
Should AI do this?
Just because it can doesn’t mean delegating it is appropriate.
What might be lost?
Would using AI remove valuable conversation, learning, creativity, judgement, discovery or team ownership?
Who remains responsible?
Who will verify the output, make the decision and remain accountable for the consequences?
Activities primarily concerned with processing information are often good candidates for AI assistance. Activities whose value comes from human interaction, judgement, discovery or ownership require much greater caution.
Many Scrum Master activities contain both. The answer doesn’t therefore have to be “AI” or “human”. AI can perform part of an activity while a human deliberately retains the parts where human involvement creates value.
In other words: accelerate the mechanics, augment the thinking, and stay human where being human creates the value.
AI Might Make the Human Scrum Master More Important
There is an interesting paradox here.
As AI becomes better at drafting, summarising, analysing information and generating ideas, the distinctly human capabilities of an excellent Scrum Master may become more valuable, not less.
Empathy. Curiosity. Judgement. Listening. Observation. Facilitation. Coaching. Building trust. Understanding context. Helping people navigate disagreement.
AI can potentially free us to spend more time doing these things.
But that won’t happen automatically.
If we use every hour AI saves simply to produce more documents, more analysis and more output, we’ve gained efficiency without necessarily becoming more effective.
The greater opportunity is to reinvest the time AI saves into higher-value human work.
That brings us back to the question we started with. The real opportunity isn’t simply asking what AI can do for a Scrum Master. It’s asking what AI can take away so that the Scrum Master has more time for the work that requires them to be human.
Perhaps, then, the measure of successful AI adoption for a Scrum Master shouldn’t be how many activities have been automated. It should be whether the Scrum Master and the team are thinking, learning and working together more effectively as a result.
So perhaps the most important question for an AI-enhanced Scrum Master isn’t:
“What can I delegate to AI?”
But:
“What is too important to delegate?”
Want to Explore Further?
For those seeking further insight, our Certified Scrum Master training now includes guidance on integrating AI effectively. Alternatively, for a more focused exploration if you are already an experienced Scrum Master, consider our half-day micro-credential course: AI for Scrum Masters.
Recent Comments