Over the last few months I keep running into articles about the death of Agile as a methodology. Some say Agile makes development more complex. Others insist the problem is not Agile, it is bad Scrum. A third group keeps arguing about Story Points, Velocity, Sprint Planning and endless retrospectives.
I think everyone is discussing the wrong problem.
While IT keeps arguing about Agile, AI is already moving development faster than we can update our processes.
The most expensive ritual in IT
Picture a normal Monday.
Sprint Planning gathers:
- Product Manager
- Project Manager
- Engineering Manager
- UX/UI Designer
- QA
- five engineers
Nine people. Ninety minutes.
Main topic of discussion:
"Is this a 3-point story or a 5?"
Someone thinks it is an 8.
An hour later the team reaches consensus.
Everyone leaves happy.
The company has just spent a few hundred — or close to a thousand euros if the CTO and CPO joined — to collectively invent a number that nobody will remember two weeks later.
Now a different picture.
I open ChatGPT. I describe the task (or, if we are already in the planning session, the product folks have described it). A few minutes later I get back:
- Problem Statement
- Scope
- User Stories
- Acceptance Criteria
- Test Cases
- SQL schema
- API description
- list of Edge Cases
- an initial technical design
And at that moment it becomes very hard for me to take the Story Points debate seriously.
Agile solved yesterday's problems
Just to be clear, I don't think Agile is bad. Quite the opposite. Agile played a huge role in shaping the industry. It pulled us out of projects that were planned for years. It brought teams closer to the business. It taught us to get feedback faster.
The problem is that many companies are stuck in 2015. They keep polishing their development processes as if the main constraint of the business were still the speed of writing code.
Spoiler. It is not.
The real bottleneck today
Over the last two years I have worked with dozens of initiatives, products and teams. And almost never was the main problem the code. The problems usually looked like this:
- nobody knows what to build;
- too many priorities;
- no owner for the decision;
- the business itself does not know what it wants;
- five executives all have the right to change direction;
- 60 initiatives sit in the backlog at the same time.
Not Scrum. Not Kanban. Not a single Agile Coach on the planet can fix this. Because these are not process problems. These are management problems.
A real conversation
A very typical meeting.
CEO:
"Why didn't we deliver this feature?"
Team:
"Because we don't have enough resources."
Ten minutes into the discussion it becomes clear:
- There are resources.
- There are engineers.
- There is a designer.
- There is QA.
The problem is somewhere else.
The company has 40 projects open at the same time and every single one of them is considered critically important. But it takes guts to tell the CEO that all their "critical right now" priorities are, frankly, nonsense. To soften it a bit, you can say: "Folks, you don't have an Agile problem. You have a math problem."
AI is changing the rules of the game
Here is what is actually interesting. The classic cycle used to look like this:
"Business → Product → Design → Engineering → QA → Release"
Today, AI is gradually shrinking every single one of these stages.
- Product writes requirements faster.
- Designers build prototypes faster.
- Engineers write code faster.
- QA gets test cases faster.
- Documentation appears automatically.
The cost of delivering a product is falling fast.
And this changes everything, because the main question now sounds different.
Not: "How quickly can we build this?" but: "How do we figure out what is actually worth building?"
The new leadership skill
I am more and more convinced that within a few years the value of a leader will not be measured by their Scrum knowledge or by the number of Agile certificates on the wall. It will be measured by their ability to:
- make decisions;
- set priorities;
- say "no";
- define direction;
- use AI as an amplifier for the team.
Because once AI starts writing code faster than people, process stops being a competitive advantage. The new competitive advantage is the quality of decisions.
My takeaway
Agile is not dead, but it is no longer the main thing. While some companies are still arguing about Story Points, others are using that same time to build a prototype, test a hypothesis and deliver the first version of a product. Not because their Agile is better - but because they understood that the game has already changed.
17 June 2026
