Berlin

November 4 & 5, 2024

New York

September 4 & 5, 2024

Will AI replace the agile coach?

Agile coaches are already using generative AI tools to help with key rituals and go deeper, quicker.
December 04, 2024

You have 1 article left to read this month before you need to register a free LeadDev.com account.

Estimated reading time: 5 minutes

While the principles of agile software development are simple, in practice, agile continues to be challenging to achieve. The complexity involved has given rise to a parallel industry of agile coaches and scrum masters to help organizations manage that challenge.

Now, as organizations explore areas where generative AI can help, will the agile coach be the next role on the chopping block? 

Agile in the age of AI

Lean and agile coach Henrik Kniberg argues in “Agile in the Age of AI” that developer teams will soon splinter off into human pairs aided by a generative AI assistant. This would result in more teams than ever.

With these smaller, human-AI hybrid teams, Kniberg predicts productivity accelerating to the point that sprints are cut down to a day or less, with the daily standup and sprint planning merging into one activity.

With that, Kniberg argues, comes increased demand for cross-team and cross-organizational coordination, which will reemphasize the need for that daily sync – and coaches to facilitate them.

However, in a recent online meetup on agile and AI, several engineering teams highlighted how they are already using generative AI as a replacement for a scrum master or agile coach, allowing it to guide agile rituals, including providing questions to ask in retrospectives.

Can GenAI play a ‘good enough’ team advisor? The news is rife with examples of generative AI offering bad advice. Yet, millions of professionals every day are relying on these tools to guide their work. It’s no longer whether to talk with a chatbot, but figuring out when you can get value from an AI conversation.

How to talk to an AI advisor

“I tend to think of [generative AI tooling] as maybe an intern or someone fresh out of college who is really eager to answer all of the questions that you ask,” said Nathen Harvey, Google Cloud developer advocate. “They want to answer them quickly and it’s that eagerness that’s really, really important.”

For first-level coaching, AI might be valuable. Just be wary of using it to drive your decision making. Especially while they are often still inaccurate. Instead, think of AI bots as tutors, Harvey said: “When we ask the right questions of a tutor or mentor, we get better responses.”

Ground your prompts in context. Instead of: List some retrospective ideas, try: You are a scrum master working on a retail app that had a two-hour outage on Black Friday. What questions should we be asking for a blameless post-mortem? Always treat it like a conversation, where you look to improve iteratively by providing more feedback as you go.

Just take its advice with a pinch of salt. Be aware that the more you know about a topic, the better you can evaluate it for accuracy. More junior colleagues will need advice and coaching on how to de-risk AI usage.

Resolving conflicts with AI tools

AI can also help coaches be more effective. “I can be much more efficient and can spontaneously design a big workshop or retro to solve problems [in response to] some team conflicts,” said Joop Nilkuha, engineering manager at A.P. Moller-Maersk, where she continues to fulfill an agile coaching role. 

Before any feature release, she and her team use AI to generate a picture of how their application will impact customers and then draft text-based feedback, asking: What could the objections be?

Recently a disagreement broke out in the form of a long asynchronous team chat around different roles and responsibilities. Nilkuha dumped the conversation into ChatGPT to get a summary of all sides of the issue, and then used other tools like Scrum Sage: Zen Edition and Scrum Master Assistant to analyze a Miro board and brainstorm solutions.

What would normally have taken her half a day was achieved in about 15 minutes. 

“The tools are very good but you can’t just use them blindly. They tend to give wrong information,” she said. Even if you ask seemingly simple things like agile and scrum principles and values, it’s more often wrong than not. “You have to be an expert in the subject to know what to use and what’s wrong.”

Generative AI as a creative partner

On the other hand, AI can help you prepare and kick off training. “AI can help accelerate our learning to an intermediate level,” said Ilona Brannen, leadership development consultant, arguing that it’s not ready to go as deep as professional coaches can.

GenAI can help you get through the beginner stage quicker, she continued, but then human interaction is required to reach the next level of collaboration, creativity, and productivity.

One benefit of generative AI is that it will better prepare people for conversations with a coach, mentor or manager. Using this tool allows you to have already worked out how to talk about what you’re trying to achieve. You can also use AI as a third party to help kick off a conversation, or as a way to brainstorm before you meet with a coach to help that session go deeper, faster. 

There’s no replacement for conversation

In the end, if a coach can spend much less time preparing, they can spend more time talking to people.

“But if we are too focused on the tools themselves, then we are not following that agile principle already,” Nilkuha warned. “We can use it to generate conversation, but if we just use the tool, we can keep people from talking to each other. And, if we don’t talk, we can’t improve.”

Generative AI may not replace these agile coaching roles any time soon, but it should be something you should be experimenting with. “It truly is thinking of this as another tool in your toolbox, another thing to bring along,” Harvey said.

For the foreseeable future, there should still be demand for agile coaches and scrum masters to guide the very real people creating software. GenAI can be useful as an outside opinion, but it shouldn’t be the only one. 

“AI as a total replacement is a bad idea,” Brannen warned, “as you would lose the emotional connection of a team, which is what a good manager and, if you are lucky, a great leader can do.”