Are we really "Post Agile"?

AI may not kill Agile, but it could kill a lot of what we’ve come to call Agile.

Think about the machinery many software organizations have built around development:

-Two-week sprints.

-Backlog refinement.

-Story pointing.

-Sprint planning.

-Capacity planning.

-Daily standups.

-Sprint reviews.

-Release planning.

Much of it exists to manage a fundamental constraint:

Software takes time to build.

Engineering capacity is scarce, so we break work into small pieces, estimate it, prioritize it, schedule it into iterations and carefully manage it through the development process.

But what happens when that constraint changes?

For teams that can take advantage of agentic development, we're starting to see software built and iterated at speeds that would have sounded ridiculous a few years ago.

That doesn't eliminate the need for disciplined product development.

It changes where the discipline needs to live.

If an agent can turn an idea into working software in hours or days instead of weeks or months, spending days estimating and managing the work starts to look a little strange.

The bottleneck moves upstream:

-What problem should we solve?

-For whom?

-Why does it matter?

-What should we try?

-How will we know if it worked?

-What did we learn?

-What should we do next?

Those are product questions, not delivery-management questions.

And they become more important when building gets easier.

This may be the biggest change AI brings to product development:

We can spend less energy managing the production of software and more energy deciding what software is worth producing.

Some Agile principles will age beautifully: small increments, rapid feedback, learning, adapting to change.

But some of the ceremonies and artifacts we've built around those principles? I'm not sure they survive agentic development.

Maybe the post-Agile product team isn't less disciplined.

Maybe it's disciplined about decisions, experiments and learning rather than sprints, stories and velocity.