AI Is Revealing the Same Dragon Agile Revealed Twenty Years Ago

Sarah Gruneisen
4 days ago
4 min read

Today was the first day of Xebia’s AI Strategy & Leadership course, facilitated by Sander Dur.

Going in, I expected to spend most of the day discussing AI.

Models.

Agents.

Automation.

Use cases.

Technology.

Instead, I found myself thinking about Agile….

That may sound strange.

After all, Agile and AI seem like completely different conversations. One emerged from software development and product thinking. The other is rapidly transforming how knowledge work is performed. Yet as the day unfolded, I kept noticing a familiar pattern.

Twenty years ago, many organizations believed Agile would solve their problems.

Today, many organizations believe AI will solve their problems.

In both cases, the technology was never the real challenge.

The challenge was us.

Throughout the day, we worked through strategic exercises designed to identify organizational problems, desired outcomes, root causes, and opportunities for AI. The exercises were deceptively simple. Rather than jumping immediately to solutions, Sander repeatedly challenged us to slow down and ask a more fundamental question:

What problem are we actually trying to solve?

Not what solution do we want.

Not what technology do we want to implement.

Not what tool seems exciting.

What problem are we actually trying to solve?

It sounds obvious.

Yet it turned out to be surprisingly difficult.

Again and again, people would identify an issue and immediately jump toward a solution. We need AI. We need automation. We need a new process. We need a reorganization. We need better reporting. We need another dashboard.

And then the conversation would circle back.

What is the actual problem?

As the room explored examples from banks, consulting organizations, product teams, sales organizations, and technology departments, something fascinating began to emerge.

We kept looking for AI opportunities.

Instead, we discovered communication failures.

We discovered assumptions being treated as facts.

We discovered conflicting incentives.

We discovered fear.

We discovered missing evidence.

We discovered organizational silos.

We discovered authority without expertise and expertise without influence.

In other words, we discovered human problems.

One of the most interesting moments for me happened during a group exercise. Our team started discussing a challenge facing consulting organizations in the age of AI. Initially, the conversation seemed to be about AI, expertise, and changing market conditions.

Yet the deeper we dug, the less it became a technology discussion.

The real questions became:

How do organizations establish authority?

How do experts communicate value?

How do leaders make good decisions when information is abundant but judgment remains scarce?

How do you translate deep expertise into language that decision makers can understand?

None of those are AI questions.

They are leadership questions.

As someone who has spent more than twenty years working with Agile, I experienced a strange sense of déjà vu.

When Agile began spreading through organizations, many people treated it as a solution rather than a mirror.

Teams adopted Scrum.

Leaders created Agile transformations.

Consultants delivered frameworks.

Organizations created Agile departments.

Yet many of the promised benefits failed to materialize.

Not because Agile was broken.

Because Agile exposed problems that were already there.

Poor communication became visible.

Weak leadership became visible.

Misaligned incentives became visible.

Siloed thinking became visible.

Lack of trust became visible.

The framework did not create those problems.

It simply made them harder to ignore.

I am beginning to suspect that AI is doing exactly the same thing.

AI is not creating most of the challenges organizations are struggling with.

It is exposing them.

A team that struggles to communicate will not magically communicate better because AI is added.

A leadership team that cannot align around priorities will not suddenly become aligned because AI writes their reports.

An organization built on poor incentives will simply accelerate those incentives.

A company with weak data practices will generate faster confusion.

A company with strong learning habits, healthy communication, and clear strategy will likely accelerate in the opposite direction.

The technology amplifies what already exists.

That stayed with me throughout the day.

One phrase that surfaced repeatedly during the exercises was the idea of slowing down to speed up. At first glance, this seems almost contradictory. Organizations are under pressure to move faster than ever. AI is often marketed as the answer to that pressure.

Yet many of the examples we discussed demonstrated the opposite.

The organizations that rush directly toward solutions often spend enormous amounts of money solving the wrong problem.

The organizations willing to spend time understanding root causes frequently move faster in the long run.

The irony is that slowing down feels uncomfortable because it feels like inaction.

It feels like analysis.

It feels like delay.

Yet rushing toward solutions often creates far more delay than curiosity ever could.

Perhaps that is why so many organizations struggle with transformation.

Technology is exciting.

Technology is visible.

Technology creates the feeling of progress.

Human work is slower.