Application Management in the Age of AI

Rethinking Ownership, Governance and Lifecycle in the AI Era

Artificial intelligence is changing how software is built, more importantly, it is changing who can build it.

Generative AI and “vibe coding” tools enable business users to create working applications with very limited technical background. What previously required a development project can now start with a prompt and an idea.

Although this will raise some new challenges, we should not treat this as something entirely new or unprecedented. Organisations have been dealing with decentralised software adoption for years. Shadow IT has always been there. Unapproved SaaS tools have found their way into enterprise landscapes for more than a decade. Low-code and no-code platforms already lowered the barrier even before AI entered the picture.

AI changes the speed and accessibility of building software, and how quickly those solutions can move from experiment to structural use.

When Software Enters the Landscape Informally

Traditionally applications entered the application landscape through relatively structured processes: business case, architecture review, implementation, handover.

In practice, that structure has never been absolute. SaaS tools were often adopted bottom-up. Teams experimented first and formalised later.

The AI era intensifies this existing dynamic:

• A team experiments.
• A prototype works.
• Colleagues start using it.
• Data flows in.
• Integrations follow.

At no clear moment does someone declare: “We are introducing a new application into the landscape.” Yet structurally, that is what has happened.

A Familiar Scenario, Now Compressed in Time

Someone in marketing builds a lightweight CRM over the weekend using AI tools.
On Monday, the team starts using it.

Within weeks:

• Customer data is stored
• Reports are generated
• Email integrations are configured
• Business decisions rely on it

The intent is usually pragmatic: solve a problem quickly.

The issue arises when a solution that began as a temporary workaround becomes a structural dependency without deliberate evaluation.

The timeline from idea to operational reliance has shortened dramatically, that compression of time is the real shift.

The Operationalization Threshold

A practical way to approach this is to define a threshold at which a tool becomes an application.

Typical signals include:

• Multiple users
• Storage of corporate or personal data
• Integration with other systems
• Use in recurring operational processes
• Dependence for decision-making

When these conditions are met, the solution is no longer merely experimental. It has entered the application landscape.

At that point, lifecycle and governance considerations become unavoidable.

A Landscape Governance Question

Once a tool starts being used structurally, fundamental questions arise just as quickly.
Who owns it? Who pays for its evolution? Who supports it when something breaks? How does it fit within the broader application landscape? What happens when the person who built it moves on?

These are the kinds of fundamental considerations that determine whether a solution can sustainably remain part of the organisation. Addressing them is part of the everyday reality of application management.

In practice, this becomes a portfolio decision. Once a solution is used structurally, organisations need to decide whether to tolerate it within clear boundaries, invest and industrialise it, migrate its functionality into a more strategic platform, or retire it altogether. The framework itself is not new. What changes in the AI era is how often these decisions need to be made and how early in the lifecycle they arise.

When organisations delay these portfolio decisions, the consequences are rarely immediate but steadily cumulative. Overlapping tools emerge. Data models diverge. Informal integrations become fragile dependencies. Key-person risk increases.

What appears as agility at first can slowly translate into structural complexity.

Deliberate governance is therefore not about slowing innovation. It is about preventing invisible debt.

An Additional Responsibility Layer

The mandate of application management remains centered around lifecycle, governance and rationalisation.

What changes in the AI era is the rate and decentralisation of software inflow.

Application management therefore needs to extend its discipline to earlier stages of the lifecycle: identifying, classifying and deliberately absorbing new solutions into the portfolio.

This does not imply absorbing all development, it implies ensuring that structural usage is matched with structural responsibility.

Embracing the Evolution

AI-driven software creation represents a significant evolution in how organisations solve problems.

It increases autonomy.
It accelerates experimentation.
It reduces the friction between idea and execution.

Those are meaningful advantages.

However, the ease of creation must be matched by clarity of ownership and lifecycle.

The easier it becomes to build software, the more deliberate organisations must be in managing what they keep, scale or retire.

Strategic Implications for Leadership

The acceleration of software creation does not merely introduce operational questions. It introduces leadership questions.

When software can emerge anywhere in the organisation, governance can no longer be positioned as a late-stage control mechanism. It becomes a design choice.

Leaders must decide how much autonomy they encourage, how explicitly ownership is assigned, and how portfolio absorption is structured. Without that intentionality, decentralised innovation gradually turns into structural fragmentation.

The tension is not between innovation and control. It is between unmanaged proliferation and deliberate integration.

AI does not weaken application management. It increases its strategic relevance. The organisations that navigate this shift successfully will not be those that restrict experimentation, but those that create clarity around what happens after experimentation succeeds.

Conclusion

AI will continue to expand the application landscape.

The real challenge for application management teams revolves around intentionally shaping that expansion so that it aligns with the already existing application landscape and it’s design principles.

Ownership, governance and lifecycle have always been central to application management. In the AI era, they become more visible and more frequently tested.

As the speed of software creation increases, the need for clarity increases with it. Clear ownership. Clear decision rights. Clear lifecycle expectations.
Application management ensures that what is built can endure.