DEV Community

Cover image for I asked ChatGPT to draft a four-hour agile restrospective - here's how it went πŸ¦ΎπŸ’¬
Pascal Thormeier
Pascal Thormeier

Posted on

 

I asked ChatGPT to draft a four-hour agile restrospective - here's how it went πŸ¦ΎπŸ’¬

Last weekend, I had the task to craft and hold a four-hour workshop for a group of people who have been working together for years. The group faced a few problems lately: Since they started working on their shared project in their spare time, motivation has decreased, and so has productivity. At some point, there wasn't any progress in months.

They wanted to identify the main problems and find solutions for them. They were very well aware that they needed to discuss a lot and take their time, hence the four-hour timeframe.

I hadn't held such a workshop in ages and was running out of time to prepare it, so I asked ChatGPT for help. We've seen ChatGPT code a lot now. Let's see if it can also act as an agile coach.

The prompt

On LinkedIn, I've seen many guides on using ChatGPT. Most of them focus on how to pose questions, but one interesting detail that keeps popping up is how people tell the AI to "pretend" to be someone else that would be very knowledgeable on the requested topic.

Grant "3blue1brown" Sanderson has shown us on Twitter what could be done with that technique:

So I tried the same. I asked ChatGPT to pretend to be a business consultant with focused experience in process optimization and agile methods.

Pretend you're an expert business consultant specializing in processes and Agile development methods.

Then I described the problem as detailed as I could and asked for a concept of a four-hour workshop:

The company you're consulting has an issue. It's a team of five people that were very motivated initially and had many ideas for bringing the company forward. They had worked hard on all their tasks and achieved a lot initially. However, since the company is a side project for all of them, they all lost motivation and started to work less and less on the actual tasks, always finding excuses. People seem demotivated now and are thinking about closing the company, even though they don't want to. They try to have regular stand-up meetings and update each other on the tasks, but often they don't have any updates. They used some agile process, which resembles Kanban. Please generate the concept for a four-hour workshop aiming at finding the reasons for the issues mentioned and finding solutions.

I used the term "company", although it's a loose group of friends working on the same project. I thought that ChatGPT could, in this particular setting, work best with a company since it pretended to be a business consultant.

Surprisingly, the concept was coherent and aimed at finding problems and generating solutions.

The workshop concept

So, ChatGPT described a workshop in six steps, essentially describing a SCRUM retrospective.

Here is a concept for a four-hour workshop that aims to address the issues you described:

  • Introduction: Begin the workshop by setting the context and explaining the purpose of the session. Let the team know that the goal is to identify the reasons for their demotivation and find solutions to get back on track.

  • Identifying the issues: In this step, the team will work together to identify the specific problems causing demotivation and decreased productivity. This can be done through a group discussion, individual brainstorming, or a combination of both. Encourage open and honest communication during this step.

  • Root cause analysis: Once the issues have been identified, the next step is to understand why these issues are occurring. This can be done through a technique such as "5 Whys" where the team asks "why" a specific problem is occurring, and then asks "why" again for each answer until the root cause is identified.

  • Brainstorming solutions: Once the root causes have been identified, the team can brainstorm potential solutions to address these issues. Encourage the team to think outside the box and consider all options.

  • Prioritizing and action planning: After the solutions have been identified, the team can prioritize which solutions to implement first. Then, they can create an action plan to put these solutions into practice.

  • Closing: Close the workshop by summarizing what was discussed, identifying next steps, and setting up a follow-up meeting to track progress and make adjustments as needed.

Throughout the workshop, it is important to focus on creating a positive and supportive environment, where everyone feels heard and respected. Encourage open communication, active participation, and a willingness to take ownership and make changes.

Although ChatGPT apparently does have some "expertise" with these topics, it does help to know agile setups and perhaps even have experience with moderating workshops in general.

I dug deeper.

That's already a strong basis. How can they set the context better in the beginning? Please generate a light and fun starting exercise for the first step.

And ChatGPT delivered.

Here's a similar exercise that can work well for small groups of 4 participants:

  • "Motivation Web": Create a large web on a whiteboard or flipchart, with a central circle labeled "Motivation" and lines radiating out to different circles labeled with various factors that can impact motivation such as "Workload", "Communication", "Recognition", "Purpose", "Autonomy", "Support", etc.

  • Give each participant a set of sticky notes, and ask them to write down their current thoughts and feelings on their motivation levels and how they relate to the different factors on the web.

  • Once they have written their notes, ask them to place their notes on the appropriate circles of the web.

  • After everyone has placed their notes, ask the group to come back together and have a discussion about the web. This will help to create a sense of connection and understanding among the team, and also give you an idea of the team's current motivation levels and which factors are impacting it.

  • This can be a great way to start the workshop, as it helps to set the context by identifying the team's current motivation levels and the factors that are impacting it, and provides a platform for open and honest communication.

  • You can use this exercise as an introduction to the workshop, and move on to the next steps of identifying the issues, root cause analysis, brainstorming solutions, and action planning.

This exercise is also light and fun, and should not take more than 15 minutes, it's meant to be a warm up that will help the team to relax and get ready for the rest of the workshop.

And this really surprised me. I didn't find any other resources on the "Motivation web" activity. So, either it found some obscure unknown method, or it came up with the activity itself, essentially showing creative skills.

The former seems more likely, though.

I asked a few more questions to elaborate on some of the activities, especially on the issue identification and the solution finding.

One of the main issues, though, was that it kept suggesting activities for large groups and had to constantly be reminded that there would only be four to five people present.

Me asking ChatGPT for an activity for smaller groups.

Holding the workshop

I mostly stuck to the script: Starting with an exercise I came up with myself, continuing with the motivation web, and then going through the issue hunt and the "five whys" activity.

It went surprisingly well. Although I had to improvise here and there, the timeframes it suggested were quite precise, and we ended up having identified some core issues and solutions for them.

However, from here on, it's our job to stick to these solutions and implement them.

I did tell the participants that I had a helping hand preparing the workshop and only told them afterwards that it was ChatGPT. Their surprised faces were priceless. Although, some felt that I was, in a way, cheating.

Conclusion

Can ChatGPT replace an agile coach? Certainly not. It lacks the emotional intelligence that is necessary to react to conflicts.

Since the chatbot is a passive entity, it can only react unless we constantly feed it with data from the daily standups and other SCRUM rituals.

Can ChatGPT support junior agile coaches, though? It darn well can! Asking for workshop concepts that are tailored to situations is a feat ChatGPT can do very well.

Generally speaking, the AI can cut the time necessary for workshop preparation a lot, giving agile coaches more opportunity to focus on other tasks.

In conclusion, I would ask it again. I would, however, take the suggestions with a grain of salt and improvise where necessary. ChatGPT can replace the initial thinking steps, but it cannot replace knowledge and experience.


I hope you enjoyed reading this article as much as I enjoyed writing it! If so, leave a ❀️! I write tech articles in my free time and like to drink coffee every once in a while.

If you want to support my efforts, you can offer me a coffee β˜• or follow me on Twitter 🐦! You can also support me directly via Paypal!

Buy me a coffee button

Top comments (6)

Collapse
 
tiagorangel2011 profile image
Tiago Rangel de Sousa

Cool article, thanks for posting! (PS: was that post made by ChatGPT? πŸ˜…)

Collapse
 
thormeier profile image
Pascal Thormeier

You're welcome, glad you liked it! Nope, this post was hand crafted, except for the quoted answers. Writing articles is too much fun to give it to an AI :D

Collapse
 
naubit profile image
Al - Naubit

Great article, you got my follow, keep writing!

Collapse
 
thormeier profile image
Pascal Thormeier

Thank you! Rest assured, I do not intend to quit writing any time soon. :D

Collapse
 
shallfer7 profile image
Charles Fernandes

Great article !

Collapse
 
thormeier profile image
Pascal Thormeier

Thank you so much! Glad you liked it!

An Animated Guide to Node.js Event Loop

Node.js doesn’t stop from running other operations because of Libuv, a C++ library responsible for the event loop and asynchronously handling tasks such as network requests, DNS resolution, file system operations, data encryption, etc.

What happens under the hood when Node.js works on tasks such as database queries? We will explore it by following this piece of code step by step.