How Jeff Sutherland is Using AI to Transform Scrum

Are you using Scrum or Agile in your projects? If so, you’ve likely experienced the time-consuming process of backlog refinement—breaking down tasks, prioritizing work, and estimating story points. So, what if AI could help streamline this process?

Download: User Story Template in MS Word format

If you’re into the world of Scrum/Agile, you’ll know who Jeff Sutherland is. Essentially, he was one of the founders of this methodology and one of the co-authors of the Scrum manifesto.

So, for me, it was very interesting to see that he’s now exploring how to use AI to enhance Scrum. In future posts, I’ll go into more detail on his approach.

As mentioned, Jeff Sutherland recently explored this idea in a LinkedIn post, sharing his experiment with using ChatGPT to manage a product backlog. Right now, there’s over 58 replies, with a lot of interesting points getting raised.

At a time when AI is reshaping industries, this raises an important question: Can AI enhance Scrum practices without undermining the human collaboration and critical thinking that make Agile methodologies successful?

Right now, there’s a lot of great points getting made on the pros and cons of letting AI run Scrum projects. Here’s five responses from the LinkedIn community that stood out for me.

1. AI is Not an All-or-Nothing Solution

Fabien Ninoles challenges the notion that AI must either handle everything or nothing at all. He argues that AI can be a valuable tool for organizing and suggesting backlog items, but the responsibility for reviewing, refining, and adopting these suggestions remains with the team. AI, in this context, is no different from tools like Jira or Kanban boards—it’s a means to an end, not an end in itself.

However, Ninoles warns of a potential danger: combining AI with “Zombie Scrum,” where teams mechanically follow processes without understanding their purpose. This could lead to disengagement and inefficiency. The key, he emphasizes, is to avoid a bureaucratic mindset where following rules becomes more important than achieving goals. AI should enhance collaboration, not replace it.

2. AI is a Predictive Tool, Not a Decision-Maker

Chris Alexander builds on Ninoles’ point by highlighting the limitations of AI. He explains that AI is excellent at predicting likely outcomes based on data but lacks the task-specific knowledge and experience required for complex decision-making. For example, breaking down work and prioritizing tasks often involves nuanced understanding and context that AI cannot replicate.

Alexander compares using general AI tools for specialized tasks to using a handsaw to drive nails—it might work, but it’s far from ideal. His takeaway? AI can assist in backlog refinement, but it should not replace the human judgment and expertise that are critical to effective Scrum practices.

3. Refinement Requires Human Conversations

Luiz Quintela pointa out that true backlog refinement involves human conversations, questions, and collaboration—elements that AI cannot replicate. He questions the value of using AI for tasks like prioritization and decomposition, which are inherently human-driven processes.

Quintela also raises concerns about data security and intellectual property, noting that Sutherland’s experiment overlooks these critical issues. While he acknowledges the potential of AI, he cautions against exaggerated claims and emphasizes the importance of maintaining human-centric practices in Scrum.

4. AI Can Streamline, But Not Replace, Human Judgment

Kamal Kumar acknowledges the potential of AI to streamline task organization and reduce planning overhead. However, he stresses that AI cannot replace the human element in understanding context, dependencies, and stakeholder alignment. The real value of product management, he argues, lies in prioritizing the right work—a skill that remains uniquely human.

Kumar’s reminds us that while AI can be a powerful assistant, it should not overshadow the critical thinking and collaboration that drive successful Scrum teams.

5. AI as a Collaborative Tool for Backlog Management

Suha Selcuk shares a practical example of how AI can enhance backlog management. His team developed an AI consultant (vdf.ai) that integrates with tools like Jira and GitHub to generate Product Backlog Items (PBIs) with high accuracy. The tool can create PBIs from a single sentence of input, significantly reducing the time spent on administrative tasks.

Selcuk’s example highlights the potential of AI to augment Scrum practices while emphasizing the need for human oversight. His work demonstrates how AI can be a collaborative tool, helping teams focus on value delivery rather than manual processes.

Download: Business Requirments Templates

Next Steps: How You Can Start Using AI in Scrum

Like I said, there were tons more comments from both Scrum and AI camps.

What I’m seeing is that the insights from Jeff Sutherland and others on LinkedIn make one thing clear: While AI has the potential to enhance Scrum practices, it’s not a substitute for human collaboration and judgment.

Over to you. If you’re ready to explore how AI can benefit your Scrum projects, here are two steps to get started:

1. Experiment with AI Tools: Begin by integrating AI tools like ChatGPT or specialized solutions (e.g., vdf.ai) into your backlog refinement process. Use it to generate task breakdowns or suggest priorities, but always review and refine its output with your team.
2. Focus on Human-Centric Practices: Use AI tools to reduce administrative overhead, but prioritize human conversations and collaboration for tasks like prioritization, dependency mapping, and stakeholder alignment.

There’s no silver bullet here, so to speak. It still takes effort to dovetail AI and Scrum/Agile practices. However, by balancing AI’s efficiency with the ‘human in the loop’ elements of Scrum, you can potentially get the best from both worlds.

Would you agree? Let’s connect on LinkedIn. Make sure to get our newsletter here.