Freshwater Futures logo

Kanban Meetup: Impact of Generative AI on Kanban and Agile coaches

It was great talking at the joint Kanban New Zealand and Melbourne Kanban Meetup’s about the impact of Generative AI on Kanban and Agile Coaching.

Untitled.png

Thanks Silke Noll and Daniel Ploeg for inviting me!

Some capabilities of AI that I demonstrated:

  • GPT-4o generated a new methodology on the fly called FluxAgile, check out the chat transcript here
  • Generated a facilitators guide for a workshop introducing Kanban, and created a questionnaire to help a team decide if Scrum of Kanban was right fit. ChatGPT transcript here.
  • Emotion: Played OpenAI’s ChatGPT 4o Voice Engine demo video here

Video of the meetup

AI summaries

To help compare AI capabilities here are some summaries from various AIs of the meetup:

Google Gemini 1.5 Pro summary

Here is an amazing summary by Gemini, and the prompt I used is here.

Short Summary

This Kanban New Zealand Meetup featured Ben Hogan, Founder of Freshwater Futures, exploring the capabilities of AI in relation to Kanban. The discussion revolved around whether AI could replace agile coaches and the implications of AI for the future of knowledge work, including potential job displacement and ethical considerations.

Full Summary

The Kanban New Zealand Meetup brought together a panel of Kanban enthusiasts, including Silka, the main organizer, and Daniel, representing Kanban Sydney and Melbourne. The featured speaker was Ben Hogan, Founder of Freshwater Futures and former organizer of the Kanban Melbourne Meetup group. Ben, now deeply involved in the world of AI, shared his insights on the potential impact of AI on the future of Kanban and knowledge work.

The core question driving the discussion was: "Can ChatGPT replace your agile coach, and can AI even replace your Kanban coach?"

The panelists explored several perspectives:

1. AI's Ability to Replicate Frameworks: Silka pointed out that AI, particularly ChatGPT, can effortlessly reproduce structured frameworks like Scrum. However, it might struggle with more context-dependent models such as Kanban, which require adaptation to specific situations.

2. AI and Complex Thinking: The discussion then shifted to AI's ability to grasp complex thinking models. Participants acknowledged the limitations of current AI in handling nuanced, context-heavy scenarios. An example was shared where ChatGPT provided low-quality answers when asked about Kanban. However, they acknowledged that AI's capabilities are rapidly evolving, with models like Gemini 1.5 from Google, capable of processing vast amounts of context, showcasing the potential for significant advancements in the future.

3. The Impact of AI on Coaching Roles: The conversation explored the potential impact of AI on the role of coaches, particularly concerning education and knowledge dissemination. The panelists agreed that AI could effectively serve as a personal tutor, offering personalized learning experiences. However, the human element of coaching, particularly in navigating complex human interactions and emotional nuances, remains crucial.

AI Demonstrations: The meetup featured practical demonstrations of AI's capabilities:

  • FluxAgile: Using GPT-4o, a new agile methodology called "FluxAgile" was generated on the fly. Check out the chat transcript here.
  • Kanban Workshop Facilitator Guide and Scrum vs. Kanban Questionnaire: Demonstrations showcased AI's ability to generate a facilitator's guide for a Kanban workshop and a questionnaire to help teams decide between Scrum and Kanban. ChatGPT transcript here.
  • Emotion in AI: The meetup highlighted the emergence of emotional capabilities in AI with a demonstration of OpenAI’s ChatGPT 4o Voice Engine. Demo video here.

Recommendations: The panelists encouraged the audience to explore AI tools like OpenRouter and Poe, which offer access to powerful language models, allowing users to experience the potential of AI firsthand.

Overall: The Kanban New Zealand Meetup provided a thought-provoking discussion on the intersection of AI and Kanban. While acknowledging the current limitations of AI, the panelists emphasized the rapid pace of AI development and its potential to revolutionize knowledge work. The meetup highlighted the importance of human expertise and emotional intelligence in coaching, even in the face of increasingly sophisticated AI tools.

Resources:

Otter AI summary

Here is a lightly edited AI summary of the meeting (from Otter.AI):

The conversation revolved around the potential of AI in Agile coaching, with speakers discussing its benefits and limitations. Ahsoka argued that AI may struggle with context and relationships, while Unknown Speaker 1 highlighted their ability to adapt to context and provide meaningful responses. Speakers also explored the potential impact of AI on Agile coaching roles, discussing the potential replacement of certain tasks by AI, while emphasizing the need for human empathy and personalized approach in smaller teams. The conversation also touched on the ethical and practical challenges of integrating AI into the workplace, and the potential of AI in various fields, including emotional intelligence, ethics, and innovation.

Action Items

  • Try more powerful AI models like GPT-4 and GPT-3 via tools like OpenAI to explore current abilities
  • Experiment giving models extensive context about projects/teams to improve answers

ChatGPT's potential to replace human coaches and consultants.

  • Dan, a meetup organizer from Melbourne, discusses the upcoming Kanban conference in Sydney and workshops with Patrick and Arlette.
  • Ben Hogan, a former meetup organizer, shares his passion for Generative AI and his company, Freshwater Futures, which aims to help people in that space.

AI projects, ChatGPT, and the reliability of information.

  • Silke questions the reliability of AI-generated content, particularly from non-experts using ChatGPT.

Powerful language models for AI writing, including GPT and Gemini 1.5 Pro.

  • Ben discusses powerful language models GPT and Gemini 1.5 Pro, with a 1 million context window.

Using AI models like ChatGPT for coaching and practice debates.

  • The speaker discusses the potential of ChatGPT and other language models to revolutionize coaching by providing valuable insights from large amounts of data.
  • The speaker encourages the group to practice debating with ChatGPT to improve the quality of their command line prompts and questions.

Using AI to generate Agile methodologies and documents.

  • Ben discusses the potential of AI to revolutionize education by providing personalized learning experiences.
  • The speaker notes that smaller AI models can provide low-quality answers, but the quality improves with focused prompts and PhD-level questions.
  • Ben discussed using AI to create a new Agile methodology and generated a LinkedIn article announcing it, (ChatGPT transcript here: https://chatgpt.com/share/bcba36f9-2094-4070-b36e-7e4365e20f8a)
  • The AI model came up with unique names and information for the new methodology, impressing the speaker.
  • Methodologist suggests ChatGPT could save time documenting work methods.

AI models' limitations in complex tasks.

  • The speaker discusses the limitations of current language models in handling complex and difficult problems.
  • The speaker suggests that while models can perform well on well-known problems, they struggle with novel situations and complex environments.

AI bias and ego in understanding machine learning models.

  • A participant mentioned they have been giving AI models difficult generation problems, but is sometimes disappointed in their responses.
  • They wonders if their own ego or biases are influencing their perception of the AI's answers.

Using language models for coaching and learning.

  • The speaker is impressed with the quality of answers given by a language model (LM) after providing examples and guidance.
  • The speaker believes that LLMs are better than humans at many tasks that they are not familiar with, perhaps marketing or sales, but not as good as experts in their domain.
  • The speaker plans to continue experimenting with LLMs and providing them with expert information to see their capabilities.

AI's impact on coaching roles, using ChatGPT for guidance and idea generation.

  • Dan reflects on AI's impact on coaching roles, seeks guidance on understanding tone of voice and approaches.

Using AI tools like GPT for project management tasks.

  • Ben uses GPT-4 to generate ideas for a workshop on Scrum and Kanban, and to create a questionnaire to help a team decide which method to use. ChatGPT transcript here: https://chatgpt.com/share/b46f7e24-8c5b-4fcf-b98f-bdef1b0801d7
  • Ben demonstrates how GPT-4 can be used to generate checklists, write a change management plan, and answer questions related to Scrum and Kanban.

Using ChatGPT for decision-making and knowledge work.

  • Dan reflects on the differences between asking ChatGPT for information and making decisions with it.

Agile coaching, AI, and team dynamics.

  • Dan discusses the limitations of human decision-making and the potential benefits of using AI in Agile coaching.
  • Silke references an article predicting the future of Agile and the role of AI in cross-functional teams.

Adapting Agile ceremonies for cross-organizational teams.

  • Unknown Speaker: Embrace adaptive planning for faster development and better alignment.

AI's emotional detection capabilities, potential for multicultural understanding.

  • The speaker is curious about ChatGPT's ability to detect emotions in real-time video and audio, and wonders if it can understand human emotion and cultural differences.
  • The speaker mentions that ChatGPT is good at multilingual emotion detection but may struggle with multicultural emotion detection depending on how the data is labeled.

AI-generated voice, emotive styles, and bedtime story.

  • Ben demonstrates AI model's ability to generate voice in various emotive styles and with a wide dynamic range.
  • OpenAI demonstrates a bedtime story about robots and love, with ChatGPT providing a dramatic robotic voice and emotional expression.

AI emotional recognition and manipulation.

  • The speaker believes AI technology is advancing rapidly, particularly in the area of emotional face reconstruction in real time.
  • The speaker raises concerns about the potential for AI to manipulate humans through techniques such as Halo bias and unscrupulous actors.

AI's potential impact on coaching and Scrum Masters, with concerns about human identity and empathy.

  • The speaker believes AI technology will augment human coaching and Scrum Master roles, but human identity and empathy will still be essential.
  • The speaker worries about the potential dangers or risks involved when AI becomes pervasive in coaching and Scrum Master roles.

AI's potential in medicine, particularly in diagnosis and treatment.

  • Unknown Speaker discusses potential benefits of AI in medicine, including accurate diagnoses and double-checking human errors.

AI's potential to enhance human senses and accelerate research.

  • The speaker believes that AI can be used to analyze data more efficiently than humans, reducing the time it takes to analyze data from 3 years to 1 week.
  • The speaker wonders how AI will impact the job satisfaction and research abilities of professionals in the field, such as oncologists and PhD students.

AI's potential in medicine, with both benefits and risks.

  • Unknown Speaker discusses potential of AI to revolutionize research, medicine, and legal industries.
  • Speaker highlights limitations and risks of relying solely on AI, including potential for errors and lack of human judgment.
Book a discovery call