If you've been an agile coach or a Scrum Master in the past four years, you've watched it happen. Coaching teams were cut. Contract roles dried up. And every week someone on LinkedIn writes the same post: AI writes the code now, so agile is over.
The first half of that is true. The jobs went away. The second half is wrong, and it's worth being exact about why.
What the layoffs measured
In a lot of organizations, the coaching role had turned into running meetings. Stand-ups, planning, reviews, retrospectives. Each one happened on schedule. Few of them changed how the team worked.
Many coaches and most Scrum Masters are missing the mark on improving team effectiveness and building truly self-managing teams. In fact, in many organizations, the Scrum Master role became barely distinguishable from a project manager (important, yes, but NOT in this situation).
When budgets got tight, someone asked a fair question: what is different because this person is here? Too often there wasn't a clear answer. No work product. No change anyone could point to. So the organization stopped paying for it.
That's hard to hear. Some of these people were never allowed to do more than run the meetings. Others drifted into tracking dates and status because that's what the organization asked them for. Either way, the market made a reasonable decision about a role that had become meetings and status reports. It did not make a decision about agile.
What the discipline actually is
Strip away the meetings and agile comes down to a few things:
- The team agrees on what to build before building it.
- The work is broken into small pieces that can be checked quickly.
- The team gets feedback often and changes course when it learns something.
- The team decides together whether the work is finished, and owns the result.
None of those is a meeting, and none of them is a status report. They are habits, and a self-managing team keeps them without being told to.
Why AI made the discipline more necessary
AI tools write code fast. They do not decide what to build, and they do not check that what they built is right. Someone still has to do both, and now it has to happen many more times a day.
A study of more than 500,000 developers found that AI tools raised commits by 240 percent but releases by only 30 percent. Even then, you have to question the value of what was released. Most of the extra code piled up waiting for someone to review it, test it, or figure out whether it was what was wanted. That pile-up is exactly what the discipline was built to prevent: small pieces, clear descriptions, fast checking, shared ownership.
It also changes who has to make the decisions. When the work moves this fast, questions about how the team works come up many times a day. A team that waits for a Scrum Master or a manager to answer them stalls. A team that manages itself keeps moving.
The teams I've seen get real speed from AI didn't drop agile. They kept the habits, dropped the parts that had turned into ritual, and made their own decisions.
What this means if you're a coach or Scrum Master
The work that's needed now isn't running the meetings or tracking the dates. It's helping a team:
- write descriptions clear enough that an AI builds the right thing the first time,
- keep review and testing from becoming the place where everything waits,
- make its own decisions about how it works, as fast as the tools now let it work.
That last one is the job most Scrum Masters stopped doing. Your measure of success is a team that needs you less over time, because it has learned to manage itself.
If you can show a team that ships more since you arrived, and explain why, you have something organizations will pay for. Show the change, not the calendar.
I'm making the full case on Tuesday, October 27: "Agile Isn't Obsolete. It's What Makes AI Work." It's live, free and an hour long, with time for your questions. Register below.