How AI is Making Agile Coaches Better
Hold on to your ergonomic office chairs. We’re talking about AI again.
To get right to the point:
I’m confident that AI isn’t going to replace coaching. I think it’s going to make it easier and more efficient by removing tasks.
I think it’s going to help us coach more.
In fact, the more I experiment with AI — the more I think it already has.
“AI isn’t going to replace jobs; it’s going to replace tasks.”
This quote is from Kai-Fu Lee, AI expert and CEO of Sinovation Ventures. It has since been paraphrased in podcasts, interviews, and medium articles around the world. When I first heard it, I didn’t know how to feel. Just in case this guy was missing something I did a deep dive: attending numerous talks, reading articles, and talking to qualified people in the world of agile about how, if, and when I would lose my job to AI.
I even asked an AI just to be safe!
Okay, if AI isn’t going to replace us, how are we going to live together? I went back to Kai-Fu Lee’s quote many times and realized there are a lot of tasks that impede the work of an agile coach.
Maybe AI can help there!
Here are four challenges myself, colleagues, mentors, and heroes in agile coaching have experienced, and four concrete ways that AI is being used to answer them.
Challenge #1 Lack of Transparency
Otherwise known as “Where in the Sam Hill is Waldo?”.
Agile coaches often struggle with gaining visibility into team dynamics, including cross-team dependencies and impediments.
This is especially true in large-scale or distributed environments where it can be challenging to track and understand the performance of multiple teams simultaneously. It’s like hunting for Waldo on a page of nothing but red and white stripes.
My first agile coaching engagement was a hard first step. I was responsible for coaching dozens of teams who were working siloed from one another. I regret to say that the majority of my first few months was spent squinting at org charts and learning team dynamics to know who to best talk to to get any idea of what the teams were facing.
Knowing what’s going on “in the field” is crucial. It’s also a very time-consuming process that means less actual coaching happens.
My first agile coaching experience was like drinking straight from the fire hose and the result is I got very little coaching done in the beginning.
The Answer: Data Analytics
AI-powered tools can analyze and interpret large volumes of data gathered in agile environments. This includes but is not limited to team metrics, burndown/up charts, velocity trends, and retrospective outcomes.
These tools can identify patterns, correlations, and anomalies in data. Coaches can use these valuable insights into team dynamics and performance in their coaching.
For example:
AI algorithms can identify recurring issues and dependencies across multiple teams in a fraction of the time it would take a coach. This lets the coach use their time to address these issues directly.
Instead of spending time memorizing org charts and team dependencies, you can immediately focus on the biggest impediment to create the most impact.
Another benefit to AI in the coaching space that I haven’t seen talked about enough is using AI to extract best practices from team data to share across the organization. This could be a huge time-saver when working in an enterprise or with teams who are slow to share with each other. It means less time prospecting and more time polishing gold nuggets for the teams.
AI-tools are already helping project managers to achieve some of the above. Project management tools like Trello, Teamwork, and Asana have embraced AI to address strategic and cyclical reporting.
Project managers and agile coaches certainly have their differences but we share similar challenges that AI is already helping us solve.
Challenge #2 Data Overload
Ever been up data creek without a paddle?
Agile coaching involves collecting and analyzing vast amounts of data. These include team metrics, feedback, and retrospective outcomes. Processing and effectively interpreting this complex data is time-consuming and at times overwhelming.
It has been proven that letting people work the way they want to makes them happier. As a fledgling agile coach, I quickly learned just how much harder that makes my job.
Dozens of teams with different tools, ways of communicating, rules about how they prioritized work….keeping it all straight was impossible.
It felt like juggling chainsaws at times.
A colleague recommended keeping physical notes which helped to save my sanity. I still have my 100 page notebook that is completely full of memos, lists, and charts for each team I worked with. And full of red for when organizational changes made it all obsolete in one fell swoop.
The Answer: Natural Language Processing (NLP)
NLP techniques enable coaches to analyze qualitative data such as meeting notes, feedback loops, and retrospective inputs. AI-powered tools can extract meaningful patterns and sentiments from these data sources. This provides crucial information to the coach regarding the team’s emotional state, engagement levels, and concerns within the team.
The real benefit here in my opinion is having the information necessary to prioritize action.
“You talk too much and it’s hurting the team.”
Back in the day pre-AI, a coach’s job involved finding a way to communicate hard-truths like the above without discouraging people from learning.
Oftentimes a gut feeling wasn’t a good enough reason to risk bruising an ego.
I have already seen NLP used in agile coaching. In our organization NLP has been used to calculate the speaking time of each member of a scrum team. This has helped provide the coach with hard data to share with the team and made it easy to track progress toward the team’s goal (in this case, reducing facilitator speaking time so that important information is shared). This was originally coded in python by a colleague who couldn’t find an answer at the time.
Vowel (https://www.vowel.com/) is just one AI-tool that looks hopeful for this context. It analyzes speaking time, generates meeting notes, and even lets you search within the recording for a specific word or subject.
With tools like this, we can confidently change “Just my opinion but…” to “Here is the data; what do we want to do?”, which I think is a much more valuable use of our coaching time.
#3 Predicting and Mitigating Risk
Is it possible to forecast your next Agile faceplant?
Identifying risks and issues early on is a fundamental goal for an agile coach in any team but most importantly when an organization is at the beginning of its agile transformation journey. Proactively anticipating these risks and taking preventative action to minimize negative impact while allowing for learning is a delicate tightrope to walk.
In my coaching engagements, I’ve often waded into unfamiliar territory where I didn’t have domain knowledge. There were a few close calls, a few near disasters. Looking back, being able to understand when to stop and when to pass GO was essential information for when a team wanted to change a workflow or process. Our attempts to mitigate this probably prevented some disastrous decisions going through but I can’t help but think…
“If we had this information from the start…”
we would have made better business decisions with less drain on energy and resources.
The Answer: Risk Assessment and Prediction
AI just may be the crystal ball we’ve been looking for. Or a piece of it anyway.
AI algorithms can help identify potential risks and predict their likelihood of occurrence based on historical data and patterns. This can enable coaches to proactively address risks, develop mitigation strategies, and prevent potential issues from derailing an agile project or transformation.
What this looks like in practice is raising a red flag to a team with a documented history of scope creep from taking on more work than they can handle. The use of predictive modeling makes it a lot easier to explain to stakeholders how their seemingly innocuous request impacts the overall goal.
Additionally, predictive modeling can be used during product ideation to help teams decide what features to prioritize.
With the use of AI in risk assessment, we don’t need to learn all our lessons the hard way, and we can skip good ideas to the front of the line sooner.
#4 Lack of Personalized Guidance
Otherwise known as “one size does NOT fit all”.
As agile coaches, we often focus on the “bird’s-eye-view” of the organization rather than nurturing one specific team out of many. In organizations with few resources, this can often mean that people or entire teams get “left behind” during an agile transformation or big organization change. Sometimes companies come back from this and sometimes the people feeling unsupported leave the company.
The majority of my coaching engagements have been in those trenches with large enterprises who wanted help with some part of their agile transformation. But one in particular sticks out to me for all the wrong reasons. I was one of very few agile coaches in a very large organization. I was unable to provide the feedback and support to everyone who needed it.
If your house wasn’t on fire, I couldn’t help you. This was due to multiple reasons, most of which I’ve already shared earlier in this article.
Fire-fighting became the norm and coaching was a luxury we couldn’t afford.
I’m not ready to make the jump to cloning myself just yet but there has to be a better way…
After attending some lightning talks and learning about AI, I found a solution! Only I don’t think people realize the potential yet.
The Answer: Intelligent Recommendations
The Agile revolution?
Out of all of the topics covered so far this is the one that I have seen getting the least amount of attention, and it’s the one I’m personally most excited about.
Using historical data analysis and continuous learning, AI-powered coaching platforms can provide real-time guidance. For example in the case of scrum, AI-guided facilitation of scrum events can detect potential issues or deviations from best practices during a scrum event.
This won’t replace the need for a coach to understand the team’s context and explain how and why. This will provide a way for teams to move forward without a coach as a permanent fixture in the team. It also removes dependence on the coach and gives the teams autonomy.
I wish something like this had existed for my past engagements.
It’s not about replacing the need for a human coach to offer context and explanation. It’s about helping teams get up and running without the coach becoming a bottleneck.
To keep with the house-on-fire analogy, I want to give teams fire prevention training and a bucket of water until I can get there with the firetruck.
So there you have it.
We’re standing on the precipice of a new era with AI kicking the door of agile coaching wide open. It can’t replace our emotional intelligence, our skill in spotting potential in people, or our ability to lead and inspire- that’s all us. But AI is a tool, a really awesome tool with a lot of potential for us agile coaches.
If things go well we get to spend less time on tasks and more time doing what matters:
Building relationships
Growing dynamic teams
Creating rock-solid organizations
Agile is about adapting to change and I see this as another iteration of the future we’re building together.
