The Sprint Didn't Get Shorter. The Backlog Item Got Smaller.

The sprint didn't get shorter. The story did. Before AI tools, one two-week sprint finished 10 stories. With AI tools, one two-week sprint finishes 30 stories. Illustration, not measured data.

People who say AI has made agile obsolete often make the same claim: sprints are too slow for AI. McKinsey's 2026 technology trends report makes a version of it, saying that work which used to move through two-week sprints can now become a continuous loop.

The claim highlights a fundamental misunderstanding of both backlog items and sprints. It mixes up two different lengths of time.

Two lengths of time

A backlog item (or user story) is one small piece of work, such as a single change a user can see. The time from starting a story to finishing it used to be a few days. With AI coding tools, I have watched teams finish a story in a few hours.

A sprint is a fixed period, usually two weeks. At the start, the team agrees on a goal. At the end, the team looks at the product increment they created and, with stakeholders, determine the right next move. Then the team meets to decide which backlog items will help them accomplish the next move.

AI tools shortened how long it takes to build a backlog item. They did not shorten the necessity to evaluate the changes made to the product and to make sound decisions about the next move.

What sets the length of a sprint

The length of a sprint was never based on how long it takes to write code. It is based on how often people can do three things well:

  1. Agree on a goal worth working toward.
  2. Look at finished work with stakeholders and decide what to do next.
  3. Change how the team works, and then see whether the change helped.

All three depend on people. Stakeholders cannot give a review their full attention every day. A team cannot tell in one day whether a change to how it works made things better.

Two weeks is a common answer. Some teams use one week, and that choice has been getting more common for twenty years. Either way, the choice is about how often people need to review and decide. It is not about how fast code is written.

What happens when we can finish backlog items faster

The team finishes more backlog items in the same sprint. A team that used to finish ten stories in two weeks might now finish thirty. The sprint is the same length. More work goes through it.

A sprint also does not stop a team from releasing software as soon as a backlog item is finished. Teams I've coached have been releasing software many times a day inside two-week sprints for more than ten years.

Why the fixed schedule matters more now

AI tools change every few months. Each change gives a team reasons to change how it works: a new way to write stories, a new way to review code, a new division of work between people and tools. If those changes happen whenever someone thinks of one, nobody can tell which changes helped. A set time every two weeks, when the team decides on changes together and checks the results of the last ones, keeps those changes deliberate.

The review at the end of the sprint also has more to cover. Thirty finished backlog items is a lot more to look at than ten. The question for the review is whether those thirty backlog items, taken together, moved the product toward the goal. If they did not, the team built three times as much of the wrong thing.

If your teams want to drop the sprint

Ask them what the sprint stops them from doing. In my client work, the answer is rarely something the sprint prevents. Then check three things:

  1. Do you measure the time to finish a backlog item separately from the length of the sprint? They are different numbers. Only the first one should be going down.
  2. Has the end-of-sprint review changed now that there is more finished work to look at? A review designed for ten stories will not cover thirty.
  3. When the team changes how it works, is the change decided at a set time and checked afterward?

Keep the sprint. Use it to review more finished work and to decide on changes to how the team works.

To see where finished work waits in your own delivery, take the Delivery Gap Assessment.