Could a Scrum GPT Challenge Scrum Masters?

The Rise of the Scrum GPT

Fewer Jobs for Scrum Masters, and now we see the universal Scrum GPT entering the competition: Will Scrum Masters change from essential practitioners to a niche role?

The job market is currently challenging for many agile practitioners, particularly Scrum Masters. Many are looking for new opportunities, while an increasing number of organizations consider the benefits they contribute to a team’s overall success. This is not just reflected in fewer job offerings for Scrum Masters; we also observe the demand for training significantly reduced.

To make the situation worse, AI has improved significantly over the last 12 months, too. Back in November 2023, OpenAI released GPTs, a hyper-customizable version of their GPT-4-based chatbot. (Please note that you need to have access to OpenAI’s paid version of ChatGPT to use GPTs.)

Besides the exciting market opportunity for many people with specific knowledge, data, or content, the question is what implication this new technology will have on the job market for agile practitioners when their roles are partly based on “knowledge” now available from a machine.

What Is a GPT?

OpenAI introduced GPTs in November 2023: “You can now create custom versions of ChatGPT that combine instructions, extra knowledge, and any combination of skills.” (Source.)

Think of them as a way to make your knowledge available to others through a customized Chatbot on top of GPT-4. While you do not provide direct access to your training material to third parties, you train the LLM with your data, information, and knowledge. For example, I just completed the Scrum Anti-Patterns Guide book, and I could use its complete text to train a GPT running on top of GPT-4. (Just ignore the legal side of this for a moment.)

Since OpenAI launched its GPT marketplace, many free offerings have appeared; just search for them:

So, I was excited to learn that Jeff Sutherland created such a Scrum GPT:

“Exciting news for Agile enthusiasts! Introducing Scrum Sage by Jeff Sutherland - a game-changer in the world of Agile. Launched to chatGPT subscribers today under “Explore GPTs” new feature. Standish Group data reveals a startling 58% of Agile teams struggle with delays, budget issues, and dissatisfied customers. Moreover, recent surveys indicate 88% of these teams use Scrum. Why? Because of widespread misconceptions about Agile, leading to poor implementation. But don’t worry! I’m here to help turn the tide, offering genuine, expert advice straight from successful teams. Let’s transform Agile practices together!”

(Source.)

Of course, I contacted Jeff and asked about the training material for the Scrum GPT: “Exciting, Jeff Sutherland! How did you train the GPT, what material did you use?”

Jeff revealed that they mainly used the second edition of his main book: “Stefan Wolpers Base knowledge is the new Second Edition of Scrum: The Art of Doing Twice the Work in Half the Time, First Principles in Scrum and the Scrum Papers. Until we got these documents as the GPT reference we would get a lot of nonsense from chatGPT. Its data includes recommendations from the 58% of Agile teams that fail.” (Source: See the previous link.)

So, I decided to put the Jeff Sutherlands GPT to the test by comparing it to my customized standard ChatGPT.

The Prompt of the Competition

Instead of asking a simple question, I used the following prompt on designing a complete workshop on Product Backlog management to test Jeff Sutherland’s Scrum GPT and my customized standard ChatGPT 4.0:

“I am planning to create a 6-hour Product Backlog management class in the form of cohort training. There will be 3 sessions of 2 hours each spread over 4 weeks to provide time for the class attendees to apply what they learned. Essential parts of the class will be creating alignment between the Scrum team and its stakeholders, creating alignment among team members, agile product management in general, and classic Product Backlog anti-patterns identified in the Scrum Anti-Patterns Guide book. Please create a concept for this class, comprising lectures but mainly exercises covering the previously mentioned topics.”

Please note: I ran the following sessions on Sunday, January 21, 2024, from 7:10 am CET to 7:45 am CET on the model GPT-4. (This requires upgrading to ChatGPT Plus.) I did not apply Grammarly to ChatGPT’s answers.

Scrum Sage by Jeff Sutherland

This is Scrum Sage’s suggestion for the workshop:

Quote:

Based on the insights from the provided materials, here’s a concept for a 6-hour Product Backlog management class structured into three 2-hour sessions spread over 4 weeks. This format will allow participants to apply what they learn between sessions. The focus of the class will be on aligning the Scrum team with stakeholders, fostering team cohesion, understanding agile product management, and addressing Product Backlog anti-patterns.

Session 1: Introduction and Creating Alignment

Objective: Introduce Product Backlog management and start creating alignment between the Scrum team and stakeholders.

• Introduction to Product Backlog:

  • Role and importance of Product Backlog in Scrum.
  • The Product Owner’s responsibilities: Deciding work, owning the Backlog, and determining the order.

• Exercise: Creating a Hypothetical Product Backlog::

  • Participants create a hypothetical Product Backlog.
  • Emphasis on refining and estimating Product Backlog items.
  • Discussion on the importance of a Definition of Done and creating value.

• Aligning With Stakeholders::

  • Lecture on the importance of stakeholder feedback in prioritizing the Product Backlog.
  • Interactive role-play exercise: Participants practice gathering feedback from stakeholders.

Homework: Participants are tasked with drafting a Product Backlog for a hypothetical project, considering stakeholder input and product vision.

Session 2: Team Alignment and Agile Product Management

Objective: Enhance team alignment and understanding of agile product management.

• Team Alignment and Communication:

  • Discussion on the necessity of constant dialogue between the Product Owner and the team.
  • Activity: Team-building exercise focused on effective communication and trust.

• Agile Product Management:

  • Lecture on agile principles in product management.
  • Case study analysis of successful agile product management.

• Exercise: Roadmap Creation:

  • Participants create a product roadmap based on their homework.
  • Emphasis on realistic, incremental planning and prioritization.

Homework: Modify and improve the Product Backlog and roadmap based on insights from the session.

Session 3: Anti-Patterns and Continuous Improvement

Objective: Identify and address common Product Backlog anti-patterns and focus on continuous improvement.

• Identifying Anti-Patterns:

  • Lecture on common Product Backlog anti-patterns.
  • Group discussion on experiences and identification of potential anti-patterns.

• Exercise: Anti-Pattern Solutions:

  • Participants work in groups to propose solutions to given anti-pattern scenarios.

• Continuous Improvement:

  • Emphasizing the iterative nature of Scrum and the Product Backlog.
  • Group activity: Identifying areas for continuous improvement in their Backlog management.

Conclusion: 

A wrap-up session summarizing key learnings, Q&A, and discussion on how to implement these practices in their work environment.

Throughout the course, emphasize the empirical nature of Scrum, the importance of adaptability, and the need for continuous improvement. Encourage participants to share their experiences and learnings between sessions, fostering a collaborative learning environment.

End of quote.

My Customized ChatGPT 4.0

This is the suggestion of my customized ChatGPT 4.0 for the workshop:

Quote:

Creating a 6-hour Product Backlog management class in a cohort training format, spread over 4 weeks with three 2-hour sessions, is a great approach to ensure participants have enough time to absorb and apply the concepts. Here’s a conceptual outline for your class, focusing on creating alignment, agile product management, and addressing classic Product Backlog anti-patterns:

Session 1: Introduction to Product Backlog and Stakeholder Alignment (Week 1)

Lecture (30 minutes)

• Introduction to Product Backlog: Basics of Product Backlog management, its role in Scrum, and its importance for project success.
• Stakeholder Mapping: Understanding different stakeholders and their impact on the Product Backlog.

Exercise (1 hour)

• Stakeholder Alignment Workshop:
  • Role Play: Participants act as different stakeholders (e.g., customers, team members, business managers) and express their expectations and concerns regarding the Product Backlog.
  • Alignment Exercise: Identify common goals and conflicting interests among stakeholders and discuss strategies to align them.

Group Discussion (30 minutes)

• Debrief and Reflect: Share insights from the role play and alignment exercise. Discuss strategies to apply in real-world scenarios.

Session 2: Team Alignment and Agile Product Management (Week 2 or 3)

Lecture (30 minutes)

• Team Dynamics in Backlog Management: Importance of team alignment in refining and prioritizing the backlog.
• Principles of Agile Product Management: Overview of Agile product management and how it influences backlog prioritization.

Exercise (1 hour)

• Backlog Refinement Simulation:
  • Group Activity: Break into small teams, each team refines a mock Product Backlog, considering different user stories, dependencies, and stakeholder inputs.
  • Prioritization Game: Teams prioritize the backlog items based on various criteria (value, complexity, stakeholder influence).

Group Discussion (30 minutes)

• Debrief and Insights: Discuss challenges faced during refinement and prioritization. Share strategies and best practices.

Session 3: Addressing Product Backlog Anti-Patterns (Week 4)

Lecture (30 minutes)

• Common Product Backlog Anti-Patterns: Overview of frequent pitfalls and mistakes in Product Backlog management as identified in the Scrum Anti-Patterns Guide book.

Exercise (1 hour)

• Anti-Pattern Identification and Resolution Workshop:
  • Case Study Analysis: Review case studies or real-life scenarios that exhibit common anti-patterns.
  • Problem-Solving Exercise: Groups identify anti-patterns in each case study and propose solutions or improvements.

Group Discussion (30 minutes)

• Sharing Solutions and Best Practices: Groups present their findings and solutions. Discuss how to apply these learnings to avoid common pitfalls in their work environments.

Additional Elements:

• Homework Assignments: Between sessions, give practical assignments where participants apply concepts to their current work.
• Feedback and Iteration: Encourage participants to share experiences and feedback during the next session.
• Interactive Tools: Use online collaboration tools for exercises and discussions (e.g., Miro, Trello).
• Guest Speakers: If possible, invite experienced Product Owners or Agile Coaches for Q&A sessions.
• Resource Kit: Provide a list of additional resources, such as articles, videos, and book excerpts for further learning.

This format ensures a balance between theoretical understanding and practical application, with a strong emphasis on interactive learning and real-world application.

End of quote.

Comparing Two Workshop Designs

The next stop was to compare both workshop designs, and I tasked both models, respectively, using the following prompt:

“I want you to compare your response with the following alternative response. I am particularly interested in the differences and similarities of both approaches to designing a workshop for a Product Backlog management cohort, the advantages one design may have over the other, and finally, which of the two workshop designs you consider superior.”

Find the following what Scrum Sage thinks about the workshop design of my customized ChatGPT and vice versa:

Scrum Sage by Jeff Sutherland Compares the Two Workshop Designs

Note: “First Design” refers to Scrum Sage’s suggestion; the “second design” to my customized ChatGPT design:

Quote:

The comparison between the two workshop designs for a 6-hour Product Backlog management class is as follows:

Similarities:

1. Structure and Duration: Both designs propose three 2-hour sessions spread over 4 weeks, allowing time for participants to apply what they learn.
2. Focus Areas: Both workshops emphasize creating alignment (with stakeholders and team members), agile product management, and addressing Product Backlog anti-patterns.
3. Interactive Exercises: Both designs incorporate practical exercises and group discussions to reinforce learning.
4. Session Themes: Each session in both designs is dedicated to a specific theme – Introduction and alignment, Team dynamics and agile management, and Addressing anti-patterns.

Differences:

1. Session Content and Flow:
   • The first design offers a more detailed breakdown of each session with specific exercises, such as creating a hypothetical Product Backlog, team-building exercises, and anti-pattern solution workshops.
   • The second design focuses more on role-playing and simulation exercises, such as Stakeholder Alignment Workshop and Backlog Refinement Simulation.
2. Methodology:
   • The first design seems to focus more on practical applications with direct tasks like creating a Product Backlog and a product roadmap.
   • The second design leans towards conceptual understanding first, followed by application through role plays and simulations.
3. Tools and Resources: The second design explicitly suggests using interactive tools like Miro or Trello and includes guest speakers and a resource kit, which are not mentioned in the first design.
4. Homework Assignments: While both designs include homework, the first design explicitly integrates the assignments into the structure of the workshop, ensuring a continuum of learning and application.

Advantages:

• First Design: Offers a more structured approach with clear tasks and objectives for each session, which might be more beneficial for learners who prefer a hands-on, practical approach to learning.
• Second Design: The inclusion of role plays, simulations, and interactive tools might offer a more engaging and varied learning experience. The addition of guest speakers can provide diverse perspectives and real-world insights.

Superior Design:

Determining which design is superior depends on the specific needs and preferences of the attendees. If the goal is to provide a practical, hands-on experience with clear objectives and tasks, the first design might be more suitable. However, if the emphasis is on engagement, varied learning styles, and a blend of theory and practice, the second design could be more effective. Both are robust in their approach and can be effective depending on the cohort’s learning style and objectives.

End of quote.

My Customized ChatGPT Compares the Two Workshop Designs:

Note: “Alternate design” refers to Scrum Sage’s suggestion; the “original design” to my customized ChatGPT design.

Quote:

Both workshop designs focus on Product Backlog management within a 6-hour class divided into three 2-hour sessions over 4 weeks. They aim to enhance participants’ understanding and skills in aligning the Scrum team with stakeholders, agile product management, and tackling Product Backlog anti-patterns. Let’s compare them based on structure, content, and pedagogical approach:

Similarities:

1. Duration and Spread: Both workshops are spread over 4 weeks with three 2-hour sessions, allowing time for practical application of concepts between sessions.
2. Focus Areas: Both emphasize aligning the Scrum team with stakeholders, understanding agile product management, and addressing Product Backlog anti-patterns.
3. Interactive Learning: Both designs incorporate interactive elements like exercises, role plays, and discussions to facilitate active learning.
4. Homework Assignments: Each session in both designs ends with homework, reinforcing learning, and encouraging practical application.

Differences:

1. Session Structure:

   • The alternative design explicitly outlines each session’s objectives, such as team alignment, stakeholder engagement, and anti-pattern identification.
   • The original design integrates these themes throughout but doesn’t specify them session by session.

2. Content Focus:

   • The alternative design appears more structured in content flow, with a clear progression from introduction to advanced topics.
   • The original design emphasizes an immersive experience, integrating theory and practice throughout the sessions without a clear linear progression.

3. Methodology:

   • The alternative design uses a more traditional educational approach with lectures, discussions, and exercises.
   • The original design might lean more towards experiential learning, though this aspect is not explicitly detailed.

4. Team Dynamics and Communication:

   • The alternative design includes a specific team-building exercise focusing on communication and trust.
   • The original design doesn’t specify such an exercise, but it could be inferred from the overall approach.

5. Addressing Anti-Patterns:

   • The alternative design has a dedicated session for anti-patterns, with practical group exercises for solutions.
   • The original design likely integrates anti-pattern discussions throughout the course.

Advantages:

• Alternative Design: Provides a clear, structured approach that may be easier for participants to follow. The explicit focus on different themes per session helps in building a comprehensive understanding gradually.
• Original Design: Likely offers a more fluid and integrated learning experience, potentially leading to deeper understanding through continuous application and reflection.

Superiority:

Determining which design is superior depends on the participants’ needs and learning preferences:

• If the participants prefer structure and clear delineation of topics, the alternative design might be more suitable.
• If they benefit from a more holistic and integrated learning approach, the original design could be more effective.

Both designs have merit and can be effective, depending on the context and the learning style of the participants.

End of quote.

Conclusion

Considering this brief Scrum GPT experiment, which took about 30 minutes, there are two questions:

1. Can you use either suggestion out of the box?
2. Which workshop design is better?

The first question has a simple answer: Of course, that is not possible. You will need to work more to turn one of the workshop designs into an actionable concept.

The second question is also not as trivial as it looks:

1. The workshop design by Jeff Sutherland’s Scrum Sage offers a clear, structured practice with specific tasks and objectives per session, making it easier for participants to follow. This method is particularly effective for learners who favor a practical, hands-on learning style, as it builds comprehensive understanding by focusing on varied themes in each session.
2. Conversely, the workshop design by my customized ChatGPT likely facilitates a more fluid, integrated learning experience that encourages deeper understanding through continuous application and reflection. Using role plays, simulations, and interactive tools enhances engagement and variety in learning, while guest speakers contribute diverse perspectives and real-world insights.

Ultimately, neither workshop design can be labeled “superior” to the other; it is a question of the use case:

• If you start with beginners, I probably would choose Jeff Sutherland’s Scrum Sage version.
• For intermediate and advanced practitioners, I would go with the design of my customized ChatGPT.

However, the interesting question looms: Do we need a full-time Scrum Master per team if ChatGPT, in general, and a Scrum GPT, in particular, are already so well-positioned to deliver some core parts of a Scrum Master’s work?

Of course, a Scrum GPT will never be able to cover the human communication and relationship building of Scrum Masters and agile coaches. But what about many other aspects of their work, for example, workshop design?

As so often, humans overestimate the impact of innovation in the short term and underestimate its importance in the long term, while GPT-5 is already in the making. At $ 20 per month, ChatGPT is a steal.