Pod Insights The Agile Daily Standup - AgileDad AI Doesn’t Eliminate Agile Teams — It Increases the Need for Great Ones - Mike Cohn About this episode AI Doesn’t Eliminate Agile Teams — It Increases the Need for Great Ones - Mike Cohn https://www.mountaingoatsoftware.com/blog/ai-doesnt-eliminate-agile-teams-it-increases-the-need-for-great-ones
Everyone today seems eager to talk about how AI is accelerating software development. Teams are shipping faster. Individuals are more productive. Entire backlogs can be written in minutes. Estimates are a click away. Code that once took days now materializes in minutes. With all this newfound speed, it’s understandable that teams and leaders start asking whether they still need the same kinds of collaboration—or even the same kinds of teams.
Yet hidden underneath all that enthusiasm is a risk that hasn’t received enough attention. In fact, I would argue it is the risk that agile leaders should be paying the closest attention to.
https://www.mountaingoatsoftware.com/blog/ai-doesnt-eliminate-agile-teams-it-increases-the-need-for-great-ones
How to connect with AgileDad:
- [website] https://www.agiledad.com/ https://www.agiledad.com/
- [instagram] https://www.instagram.com/agile_coach/ https://www.instagram.com/agile_coach/
- [facebook] https://www.facebook.com/RealAgileDad/ https://www.facebook.com/RealAgileDad/
- [Linkedin] https://www.linkedin.com/in/leehenson/ https://www.linkedin.com/in/leehenson/
Transcript 0:15 Speaker 00 I know it's hard to believe, but here we are. It's 2026 and I have a theme going this week where it feels like everybody thinks, "Agile's done, Scrum is home, and protect this all from the evil that has come." It's just been pretty insane. And one of the things that I wanted to do to help break it up a little bit because I was getting a little burnt out just saying, "I want to read a letter from my dear friend Mike Kohn. Now, Mike and I have known each other for a number of years. We're both certified Scrum trainers. We're both absolutely amazing people." And I love Mike. It's no secret. And he sends me emails each week. I want to read this one. This is an article that he published. So I'm going to review his article.
0:56 Speaker 00 The article was recently published. It says, "AI Doesn't Eliminate Agile Teams, It Increases the Need for Great Ones."
1:03 Speaker 00 I actually wholeheartedly agree with this sentiment. Let's see where Mike takes it. "Everyone today seems eager to talk about how AI accelerating software development. Teams are shipping faster, individuals are more productive, entire backlogs can be written in minutes. Estimates are a click away. Code that once today's now materializes in minutes. With all this newfound speed, it's understandable that teams and leaders start asking whether they still need the same kinds of collaboration, even the same kinds of teams. Yet hidden underneath all that enthusiasm is a risk that hasn't received enough attention. In fact, I would argue it is the risk that agile leaders should be paying the closest attention to. AI does not eliminate teams. It increases the need for great ones. Hard stop."
1:49 Speaker 00 Yes, yes, 100% yes. I think that I'm seeing the very same thing. I'm seeing so much excitement around AI and so much energy and so much positive energy and so many wonderful things happening. But I do see this growing concern. The faster a team moves, the more collaboration it requires. The more productive individual contributors need to become, the more their work needs to be better coordinated. And the more powerful AI becomes as a tool, the more dangerous it is when someone uses it in isolation, in a silo, in a bubble. Put simply, AI amplifies everything, including misalignment and misfortune.
2:27 Speaker 00 Unfortunately, many teams are responding to AI by collaborating less, and that's where things begin to fall apart. And I agree. One of the things that I said in the early days of scaling, when scaled agile came out and all the others, is that scaling is great if you have something good to scale, but if you're scaling crap, you're going to get scrap.
2:48 Speaker 00 That was weird. It just flowed so well. Anyway. Over the past few years, I've watched teams adopt AI tools in ways that unintentionally reduce communication. Developers can now accomplish huge amounts of work on their own. Product owners can use AI to generate split back and generate and split backlog items, or even prioritize a whole roadmap. Teams start to believe they no longer need to talk, especially not as often, because the tooling is doing so much more to work. The theory goes if AI can generate stories, write code and test, perhaps the team can get by with fewer discussions. The reality is a very different story. I've seen teams that, excited by AI's efficiency, quietly slip into working as a collection of individuals
3:32 Speaker 00 rather than a cohesive unit. They may hold the same meetings, but the communication within those meetings becomes thin and surface level. Outside the meetings, collaboration dwindles even further. Let's stop for a second. This is also very true. I'm seeing many of the same things. If anything, one of the things that I've been pushing a lot recently is something I can said back in 2002. He said, quote, "No, not one line of code should ever leave the building without at least having two or more sets of eyes take a look at it." The concept of peer review, pairing, swarming, grouping together, mobbing, whatever you want to call it, right? What I'm seeing is, in order to apply AI effectively, it does require governance and human oversight. And I think that people are missing
4:18 Speaker 00 that step. They feel like, we're being hyperproductive. And once again, the focus is on output, where they're focused on delivery, and they're not focusing on outcome, which is going to yield better results, happy customers. So I think it's just this, this bridge that we still need to cross from output to outcome with where our focus is. And I think that's going to help us regain some of that pairing, peering mentality. And it's going to help us see better integration, better code creation. It's going to help everything along. We're going to end up with better products and services.
4:52 Speaker 00 Here's an example. Consider one team to dramatically accelerate development because AI allowed us to refine and build features faster than ever. Impressed with the speed, they assumed their longstanding collaboration habits were no longer necessary. They started talking less, they started sharing less, and they definitely coordinated less. And predictably, they began building a series of features that did not integrate well and didn't match stakeholder expectations. Ding ding ding. They had increased their output, but degraded their outcome.
5:23 Speaker 00 I read it. I said it first, then I read it, Mike. This is what's funny about us. You know what's you know what's crazy? Sometimes we finish each other's sandwiches. That's what I was gonna say. I've never met someone who thinks so much like me. Jinx. Jinx again. Yeah, we finish each other's sentences all the time. That's why I love you, man. Absolutely. Absolutely. AI is not just speeding up development. It's reshaping what the key roles inside the team actually do, and these changes are significant. And understanding them helps us understand why collaboration remains so important. The problem here wasn't AI. Their problem was collaboration, and AI magnified their lack of collaboration.
6:09 Speaker 00 It's just a fact. Just a fact. Product owners may be the biggest beneficiaries of AI in the short term. AI can help them write proper backlog items, generate acceptance criteria using Gertan, analyze customer sentiment, summarize research, even propose priorities. For a role that's always struggled with having enough time to do the job, this is a very welcome change. But with that support comes temptation. It becomes too easy to delegate too much thinking to AI. A product owner who lets AI drive backlog prioritization or roadmap decisions risks drifting away from the customer needs, business realities, and strategic intent. AI can assist, but it cannot own core decisions such as vision, marketing position, customer segmentation, or business
6:50 Speaker 00 models. AI is a powerful assistant, but shows and proves to be a poor navigator.
6:58 Speaker 00 Nothing can replace the concept of BAF A Ta. Business analyst, functional analyst, technical analyst to help the product owner do their thing. And Mike, I agree again, wholeheartedly with what you got here. This is something that has been an ongoing problem, and I'm glad that we see eye to eye on this. I've seen product owners who let I do most of the heavy lifting, only to discover later that the product has wandered off course. It's an example of how AI can be a powerful assistant, but a poor navigator. For years, I believe Scrum Master role would inevitably shift as teams become more experienced. As a team becomes more proficient with Scrum, team members absorb some responsibilities as a Scrum Master. AI is definitely accelerating that shift.
7:39 Speaker 00 The future Scrum Master is less of a process guardian and more of a human dynamics expert. They will likely work across teams, operate part-time on any single one, and spend far less energy worrying about prescriptive frameworks. Instead, they will focus on helping people work well together, asking tough questions, facilitating constructive disagreements, and guiding the team through difficult decisions. These human-centric skills will define the role. Scrum Masters and Agile coaches who clean the checklist and frameworks will become less effective. Those who cultivate emotional intelligence, facilitation skills, and the ability to coach collaboration will become essential. I've seen Scrum Masters who remain overly focused on rituals and mechanics
8:21 Speaker 00 while their teams quietly slide into dysfunction and dysfunctional communication patterns. In an AI-driven world, that approach will not work, and that is also true.
8:31 Speaker 00 Mike, one of the things that I discovered here is that Scrum Masters are starting to be asked to do more change management responsibilities within the organization. Organizations want someone who's a fearless leader, a coach, someone who can lead them through organizational change. They're not just looking for someone who can be a meeting jockey or a Jira jockey. They want someone who, you know, in the past, you know, was really good with one team. They want them to expand that knowledge across two teams, three teams, and then they want them to help build that Agile Center of Excellence. There's so many additional responsibilities that they need, and I think that we're starting to see those shifts, which is pretty incredible. And then for developers, of course, developers have also experienced extraordinary leaves in productivity. AI helped them generate code, uncover bugs, explore architectural patterns and write tests.
9:14 Speaker 00 This enables them to accomplish far more in less time. And with an increased capacity comes a shift in team design. Many products that previously required teams of eight members might now be able to be built by a team of four or five. This shift also changes how teams think about specialization. We still need world-class experts with highly specialized problem-solving skills, but many teams will rely on strong generalists who possess broad skills and deep AI fluency. This is where AI is digging in. It also means that developers, moving faster than ever, must communicate more frequently to avoid stepping into each other's work. And while AI accelerates their capabilities, human alignment ensures those capabilities do not become chaos-driven. And there's nothing more true than that.
9:58 Speaker 00 Developers are focusing more on quality management. Developers are focusing more on integration. I've said this for the last several months, and now we have the supporting evidence. All of these changes point toward a future where agile teams look different than they did even just a few years ago. I expect teams to become smaller and more fluid. In many organizations, team membership will include a stable core supplemented by temporary specialists, who join for a few weeks or months to work on highly complex tasks before moving on, hard stop. This is where I talked about having more of a Scrum-Bon facilitation pattern, where you had people operating as specialists in Kanban, doing more services roles, where they drop in, do some work and pull out,
10:41 Speaker 00 as opposed to having, you know, teams of consistent individuals that stayed that way forever. You would have a core team, but you have other people who would step in. So that's in line with what we've been talking about. Teams will also continue to become more globally distributed. The work-from-anywhere movement has only strengthened, and AI will make that collaboration across time zones even easier. Facts. Yet, despite these changes, the definition of a team should remain simple. A team is a group of identifiable people pursuing a shared goal. The definition still holds, even when membership becomes more fluid. As AI lowers the cost of experimentation, teams will attempt more ideas, pursue more niche products and pivot more quickly. This makes leadership more, not less, important.
11:24 Speaker 00 Leaders must provide direction, so experimentation does not dissolve into chaos. They must set strategic intent, define boundaries and articulate a north star that teams can use as an anchor. Since, putting it all framed up, since AI accelerates work, the practices that hold teams together become even more essential. That's where we need to give our focus. Sprint reviews are critical alignment moments. As AI speeds up delivery, ensuring alignment between what the team builds and what stakeholders expectations are, only through real human feedback and the interpretation of that feedback can teams ensure that AI accelerated work remains on target. We need to have those conversations.
12:06 Speaker 00 Backlog refinement is another practice that we must, that must remain human-led. AI can draft backlog items or propose ways to split stories, but only humans can decide if those items contain the appropriate amount of detail, whether the intent is right, and whether each aligns with the product vision. Daily communication becomes more essential as teams move faster. That communication doesn't need to be in a formal meeting. Many teams will choose asynchronous check-ins rather than traditional daily scrums. But the key are the frequencies and the consistency. When a team can build features in hours, they need to talk about what they're doing more often in minutes, not less. Sprint planning or iteration planning is similarly indispensable.
12:49 Speaker 00 Even if AI system forecasts, the team still must align on intent and understanding of how the work fits together. Planning keeps them grounded in a shared understanding. AI accelerates the work? Collaboration prevents accelerated mistakes. We don't want to scale crap. We want to scale good things. Leaders are often asked how they know AI is helping or hurting their teams. The answer doesn't require a dashboard full of metrics, just ask the question. Are team members communicating more frequently than they did before they used AI? If communication is decreasing, collaboration is weakening, and the teams are heading into the danger zone. The faster the team moves, the more essential communication becomes. I'm going to pause there because I feel like this is an incredible article.
13:34 Speaker 00 I'll drop a link down there. And thank you, Mike, for sending these wonderful emails. If you have a topic you want to discuss with me, or Mike, or Mountain Goat, or Agile Dad, feel free to reach out to me. Learn more at agiledad.com and we'll make that happen. And as always, we encourage you to stay well, stay healthy, and stay agile, my friends. Until next time, do take care. We'll see you soon.
Previous Scrum Is NOT Dead... It’s Obsolete? Next Agile in 2026 Looks Nothing Like Scrum Page updated November 6, 2026 · transcript and summary are machine-generated