Is Chat GPT Going to take over Agile Coaching?
Update:
A year on, this answer remains the same, and I have written further on whether ChatGPT is ready to lead your team.
Chat GPT has taken the internet by storm in the past couple of months, and even before it’s emergence, there has been a lot of conversation in the agile coaching community about Artificial Intelligence (AI) coming for our jobs (as with many industries). Since the emergence of Chat GPT it has been called everything from a sophisticated parlour trick, the google replacement, the Stack Overflow (question and answer website for developers) replacement, or our new computer overlord. Does it know agile, though, and is this the first step in AI coming for our agile coaching jobs?
In a word, yes, Chat GPT does know a whole lot about doing agile, in other words, the application of frameworks and processes for the delivery of value. Chat GPT is now a certified scrum master and has even interviewed for its first role. It even gives a really good answer about what it means to “be agile” vs. “do agile”. We don’t need to run for the hills yet though, as Chat GPT only knows the data that it has been fed and does not have the ability to think outside the square, as a human would. It does leave the question about whether is can coach an agile team. I thought I would let it speak for itself, so I asked it to write a blog post for me about whether it will take over agile coaching jobs. Here’s what it had to say:
As technology continues to advance, many industries are beginning to incorporate artificial intelligence (AI) into their workflows. One area where AI has the potential to make a significant impact is in agile coaching. But, will Chat GPT and other AI take over the jobs of human agile coaches?
On one hand, AI has the potential to automate many of the tasks that agile coaches currently perform. For example, Chat GPT can be used to generate reports, facilitate meetings, and even provide guidance on agile best practices. This could lead to a reduction in the need for human agile coaches.
On the other hand, there are certain aspects of agile coaching that AI may not be able to replicate. Agile coaching is not only about providing guidance and advice, but also about facilitating communication and collaboration among team members. It's also about coaching individuals and teams to be self-sufficient and to continuously improve. Human coaches are able to read the room, sense the team's mood, and adapt the coaching accordingly. AI is not yet able to do that.
Moreover, Agile is about people, it's about leading and facilitating change, and helping individuals and teams to adapt to new ways of working. Agile coaches are experts in dealing with human behavior and team dynamics, which is something that AI is not yet capable of replicating.
In conclusion, Chat GPT and other AI technologies do have the potential to automate certain tasks that are currently performed by agile coaches. However, they are not yet able to replace the human element of agile coaching. AI can assist and facilitate the work of the coaches but it cannot replace the experience, the intuition, and the ability to understand and work with people that a human coach has.
Hmmm, okay…. Sounds like a firm “maybe” to me, but how good is it really? It turns out, pretty good, but with limitations, including one huge one. I decided to dig a little deeper and try to find out what it knows. I asked what it thought an agile coach does, and it gave a pretty good, text book answer:
An Agile Coach is a professional who helps organizations and teams adopt and improve their use of Agile methodologies, such as Scrum, Kanban, and Lean. Agile Coaches work with teams to improve their processes, increase their efficiency and effectiveness, and help them deliver value to customers. They may also work with leaders and managers to help them understand and support the Agile process, and to help them create an environment that is conducive to Agile success. Agile Coaches may also help teams identify and overcome obstacles that are preventing them from achieving their goals, and they may help teams develop better communication and collaboration skills.
Not particularly strategic, but pretty good. Chat GPT is a Language Learning Model (LLM) that is fed large amounts of data, or as it calls it, a corpus. It goes through a process of training based on this corpus and input/output expectations of human text. Basically, it predicts what answers it should give, based on the corpus of data and patterns it has been taught. I asked it what in the corpus it learned about agile coaching from, and it provided me with this response:
I have been trained on a wide variety of texts, some of the books that are considered as a valuable resources for Agile coaching are:
"Scrum: The Art of Doing Twice the Work in Half the Time" by Jeff Sutherland. This book provides a comprehensive introduction to the Scrum framework, and is considered one of the best books on Agile coaching.
"Crucial Conversations: Tools for Talking When Stakes Are High" by Kerry Patterson, Joseph Grenny, Ron McMillan, and Al Switzler. This book provides guidance on how to have difficult conversations in a constructive way, which is an important skill for Agile coaches.
"The Lean Mindset: Ask the Right Questions" by Mary Poppendieck. This book provides a comprehensive introduction to Lean thinking and its application in software development, it also provides guidance on how to develop a Lean mindset, which is important for Agile coaches.
"Scrum Mastery: From Good to Great Servant-Leadership" by Geoff Watts. This book provides guidance on how to be an effective Scrum Master and Agile Coach, and covers a wide range of topics, from facilitating ceremonies to coaching teams.
"Coaching Agile Teams: A Companion for ScrumMasters, Agile Coaches, and Project Managers in Transition" by Lyssa Adkins. This book provides guidance on how to coach Agile teams and help them achieve their full potential, and covers a wide range of topics, from developing a coaching mindset to providing feedback.
These are just a few examples of books that are considered valuable for Agile coaching, there are many more resources available, including other books, articles, and online resources that can provide further information on Agile coaching and Agile methodologies.
Great reading list, and it did say that it has learned specific methodologies from “Scrum Alliance, Scrum.org, Agile Alliance, and other reputable organizations that provide training and certification in Agile methodologies”. Not bad, so its knowledge is very much theory based, and does not at all rely on any real world experience in implementing these frameworks, other than what it has been told. This is a big strike, in my opinion, as one of the most important things an agile coach must have is first hand experience in being an agile practitioner. How else can you work with teams and organisations to apply patterns specific to their context? Which brings us to the final question of our new AI friend; what would you need to learn to be able to fully coach agile teams on your own:
To be able to coach Agile teams on my own, as a language model, I would need the following capabilities:
1. Knowledge of Agile methodologies: A deep understanding of Agile methodologies such as Scrum, Kanban, and Lean, as well as the principles and practices of Agile development.
2. Communication skills: The ability to communicate effectively with team members and stakeholders, and to facilitate conversations and discussions in a way that promotes collaboration and understanding.
3. Problem-solving skills: The ability to identify and solve problems that may be preventing the team from achieving their goals.
4. Adaptability: The ability to adapt to different teams and different situations, and to adjust my coaching style as needed to meet the team's specific needs.
5. Empathy: The ability to understand the team's perspective and to work with them in a way that promotes trust and respect.
6. Facilitation skills: The ability to facilitate Agile ceremonies such as sprint planning, daily stand-ups, retrospectives, and demos effectively.
7. Leadership skills: The ability to lead by example and to inspire and motivate the team to achieve their goals.
8. Coaching and mentoring skills: The ability to coach and mentor team members, and to provide guidance and feedback that is tailored to the team's specific needs.
9. Continuous learning: The ability to continuously learn and keep up to date with the latest developments and best practices in Agile coaching.
10. Human-like capabilities: The ability to sense the team's mood and emotions, to understand the team's motivation and demotivation, and to act accordingly.
As a language model, I have knowledge about Agile methodologies, principles and practices, but I lack the human-like capabilities such as sensing and understanding emotions and motivation, which are crucial for an Agile coach to effectively work with a team. It's important to note that even if I were to possess all the capabilities listed above, I wouldn't be able to replace a human Agile coach as the human experience and ability to understand and react to the human emotions and motivations is crucial for success of the coaching.
And that’s the critical, final (only one needed?) nail in the coffin. The textbook, framework based knowledge is, in my opinion, the easiest thing about what we, as agile coaches, do. The most important skill, and one that still requires humanity, is the ability to empathise with people undergoing a change, and to be able to pragmatically apply concepts to their specific context to help them to solve problems. So, no, Chat GPT, nor any other AI as we know it today, is going to come close to taking any job that has such a fundamental human thing at its core. And, let’s just ignore the fact that it said it needs coaching skills, lol.
As a parlour trick, or just to check it out, sign up for a free account here, and ask it anything: https://chat.openai.com/chat
I will be continuing to explore its use in augmenting our work as agile coaches. As a knowledge resource, it’s pretty stunning, and it can even code! Take for instance a use case of helping a development team to apply Test Driven Development for the first time in the image below. Human emotion though? A long way off.
