If I drew you a chart that mapped the trajectory of the popularity of the terms “agile” and “AI” it would look like a strand of DNA. The agile curve would be plummeting while the AI line would be skyrocketing up and to the right. There’s a new kid in town and the “old stuff” is a lot less interesting. But wait just a second. Just because something’s been around a while and there’s a shiny new object to adore, does that mean the old thing is useless? I don’t think agile is done just yet. 

What’s in a name?

After more than 20 years of use the word “agile” evokes a variety of emotions and reactions in those who hear it. I would guess that for a lot of folks that association isn’t positive. A lot money was spent on a lot of “transformations” that yielded a whole lot of nothing in the long run except maybe some new job titles and a new vocabulary. That baggage is real and it risks dragging the benefits of agility down with it. As organizations begin to wrestle with AI, throwing away good ideas simply because they’ve been around a while risks repeating the same mistakes of the past. That said, if the word “agile” instantly evokes resistance in your organization, you have my permission to never say it again. 

The ideas of agility are more applicable than ever

Agile was conceived as a response to the complexity of software development. This was in the days of deterministic systems. Today’s AI-powered systems are non-deterministic increasing the complexity of the work 1000 fold. Agile ways of working — small chunks of work, short cycles, inspecting, adapting, continuous improvement and learning — make even more sense in a world with increased uncertainty in software development. 

The way we build software may be evolving but so is the risk associated with that work. How do we know we’re building something of value? How do we ensure the system produces useful output for our customers? The push from corporate leaders to “just add AI” is bringing us right back to the days of the mobile web (“just build an app”), Web3 (“just add blockchain”). We’re blindly building and shipping and hoping it’ll be valuable for the sake of ticking a box. 

Agile deals with this head on. It admits the complexity and breaks it down into manageable chunks that can be tested, validated, optimized and continuously deployed. And if something doesn’t meet our expectations we adjust course. That is agility and that is the promise of being agile in the AI age — building less AI crap. 

Keep the work, change the language

If you’ve dedicated a chunk of your career to using agile ways of working, rest assured that time wasn’t wasted. The methods hold and still make sense. What might have to change are the words you use to describe this work. If your team or org is allergic to “agile” phrases and names, change them. Who cares? As long as the fundamentals remain in tact, the agility of the organization will remain in tact. And the AI products you end up building will stand a better chance of success.