Podcast: Misconceptions around Agile in an AI world

In this week’s episode of our podcast, What the Dev?, we spoke with David Ross, Agile evangelist for Miro, about some of the misconceptions people have about Agile today, and also how Agile has evolved since its early days.

Here is an edited and abridged version of that conversation:

Where do you see the change from people doing Agile and thinking they understood it, to now? What do they have to take into consideration for this new modern era?

I have been in software development for almost 20 years, and it’s been an interesting evolution for me to watch what Agile meant maybe 15-20 years ago versus how it’s perceived today. I just remember back in the early days of some of the very first Agile transformations that I was part of, it was very much all about following a process and having fealty to specific frameworks, be it Scrum or Kanban or whatever the case might be. And the closer you were to perfection by following those frameworks, the closer you were to God, as it were, like the more Agile you could claim to be. 

And what we forgot in all of that was, of course, that the Agile values and principles don’t prescribe any particular framework or approach. You’re supposed to put people and interactions over tools and processes. Well, if you are enforcing processes and you’re asking people to interact via tools, that kind of defeats a lot of the very fundamental sort of values of Agile right from the get go.

We also have problems, in that a lot of people came into the industry, and maybe people who were not sufficiently trained or had enough experience in real, good Agile practices, and there was just a lot of bad, bad Agile out there. You know, people who got a two-day certificate stamped and said, hey, I’m going to come in and now enforce Scrum processes on this team and coach them to higher levels of agility, and that’s not a recipe for success.

This has been true of DevOps, value stream management, you you, these are just vague, non-prescriptive processes to follow. But nobody says you have to be doing X, Y and Z to be Agile, or be doing full DevOps, or be doing value stream. It’s kind of like, well, we’re just going to leave it up to you, adopt what you want, throw out what you don’t want, we don’t mean to be prescriptive. But, I think that has added to so much confusion in these markets over the years. So where we’re at now, and you’re talking about evolving into this modern era, what’s impacting it? Is it simply cloud-native computing? Is it AI? Is it all of the above? 

I feel like Agile reached this sort of peak, where people were finding that they weren’t really getting the value that had been promised as a part of an Agile transformation. They weren’t seeing the value for their customers, they weren’t seeing their value for their teams. And, you know, the house of cards started to fall apart a little bit. And let’s be honest as well, one of the things about Agile was you had to have co-located teams, so that’s one sacred cow that got sacrificed during Covid, because co-located teams just wasn’t a possibility, and we’re not in that world anymore. 

And honestly, from where I sit, Agile was invented to solve a very specific, defined problem within software development, which was software development delivery and making sure that you weren’t constantly missing deadlines, and that you were delivering the right level of value. And I think a lot of those problems have kind of been solved, and Agile has kind of expanded beyond the boundaries of just software development as well. And people are kind of seeing that it’s not one size fits all. It needs to be more adaptive. It needs to be more pragmatic and less prescriptive. 

And so that’s kind of where we are right now. I feel like where we’re in a period of retrenchment and reinvention of Agile. People are starting to see that prescriptive frameworks just aren’t going to work for them. And a lot of the customers that I talk to are evolving and coming up with their own sort of custom approach. And they’re maybe using different vocabulary, different language, but they’re still doing things that are Agile, but they’re just not recognizable to somebody 10-15 years ago.

You bring in cloud-native computing, where now you have a whole lot of moving parts, where it isn’t just a monolithic code base going through, but you’re calling APIs, you’re using Kubernetes, containers. And all of these complexities kind of change the looks of things, so how do those things affect the way that people have been doing Agile, and what adjustments have they had to make for those types of things?

I think they’ve kind of stepped away from prescriptive frameworks, and many times they’re just adapting. This is really, honestly what they should have been doing all along. You should have not been prescriptive, you should have been able to adapt your processes, and even if it’s not pure to the framework that you started with, it’s okay for you to move in that direction. So people are, I think, moving away from those defined roles that were part of those frameworks. I think that that’s probably a good thing. Rather than, you know, you’re a product owner or you’re a Scrum master, or all of those kinds of things, moving away from prescriptive titles I think is one thing that I’ve seen them do.

Also, working with tool sets that are less rigid and more flexible. So if you are trying to run everything within a very defined set of tools, and those tools define your workflow, that’s very constrictive, I feel like for a lot of a lot of companies and a lot of teams, and they’re trying to find a better way to organize themselves and to support their ways of working using more flexible tool sets.

How is AI impacting Agile development?

Well, you know, I would be lying if I could say that anybody knows the answer to that, right? We’re still in the very early days of that revolution. But some things that I can kind of see on the horizon as potential outcomes and potential impacts of AI are is it going to affect the team size? If you think about an Agile team generally, they used to prescribe that the ideal size is six plus or minus three, and you have to have these specific skill sets on it. Maybe team sizes are going to shrink a little, and you’re going to have maybe one or two developers on a team, and then they can orchestrate a series of AI agents that do a lot of the work that other specialists would have done in the past, like QA or specific database tasks or things like that. So I definitely think it’s going to affect the team composition, the team structure, and the team size. 

The other thing that I think it’s going to really impact as well is a lot of the monotony of some of the tasks that get done are probably going to be taken over by AI. And you see that across all industries, right? What does that mean? It means that it’s going to free up the really talented people on Agile teams to do sort of those higher level strategic thinking. You know, the things that AI can’t do yet. Maybe it’ll do it one day, but it can’t do it today where it’s thinking strategically and thinking about human dimensions of what they’re building and making sure that it’s being guided in that direction. The actual coding work or testing work will probably be taken over by some form of an AI, but we are going to have the ability to focus our efforts on those higher order or higher complexity activities. 

So you really have to prepare yourself individually. You have to bring your skill set up, and you also have to know how to work with an AI, because if those AIs are going to be your assistants, or they’re going to be an embedded part of your team, you have to know how to be able to orchestrate and run a series of AI agents that are going to get the work done that other human beings would have done before. So I really think that’s going to happen. What does that mean for Scrum masters specifically? I think Scrum masters, again, will have to evolve in a different direction and focus more on the human element. We’ve always said that Scrum masters are also Agile coaches, but we haven’t really taken that to heart. And I feel like that’s something that Scrum masters really need to embrace in this new era of being able to coach human beings and have high emotional intelligence. AI doesn’t have emotional intelligence. We do. So we need to be able to make sure that the human beings on our team are supported and have what they need to collaborate and to be successful, and then leave the drudgery to the AI.

The post Podcast: Misconceptions around Agile in an AI world appeared first on SD Times.

Source: Read More