What does that mean specifically?

Change costs in an enterprise context consist of:

• data migration
• integrations
• customization
• Training and change
• contract logic

When agents read APIs, match schemas, generate mapping suggestions, and orchestrate ETL processes, at least part of this effort is significantly reduced.

Especially in the CRM and ITSM environment.

Providers such as Salesforce and ServiceNow have historically benefited greatly from structural barriers to change.

ERP platforms such as SAP or Oracle are even more robust; processes, compliance, and financial logic are deeply embedded in their core.

But even there, pressure is building.

Why is the market reacting nervously?

At the beginning of 2026, software titles were sold off massively.
Narrative: "AI eats SaaS."

Whether this is exaggerated remains to be seen.

Something else is interesting:
The market is beginning to reassess switching costs.

And that is rational.

When agents halve the migration hurdle, bargaining power shifts.
Not tomorrow.
But structurally.

Technical reality; One Click is marketing

As someone who has been involved in ERP, CRM, and data migrations for decades;

The hard work is not in exporting CSV.

It is included in:

• semantics
• referential integrity
• historically developed workarounds
• Role and authorization models
• regulatory requirements

Schema matching has been a subject of research for 20 years.

Agents can make suggestions.
Responsibility remains with humans.

Governance is not a feature; governance is architecture.

Regulatory tailwind

The EU Data Act will make cloud switching easier from a regulatory perspective.
Portability is politically desirable.

That is to say:
Contractual hurdles are decreasing.

Now that technical hurdles are also being lowered by AI, a new dynamic is emerging.

Not AI alone.

My clear position

SaaS is not dying.

But the business model is shifting.

• Away from Seat Logic
• towards outcome
• away from lock-in
• towards data and process intelligence

The winners will be platforms that:

• build semantic data layers
• Integrate agents in a controllable manner
• Solving compliance cleanly
• offer genuine process sovereignty

Providers who only deliver superficial data will lose out.

And what does that mean for you?

If you are a CIO, CDO, or managing director;

Don't ask yourself whether AI will replace SaaS.

Ask yourselves the question;

Is your data structured in such a way that an agent could migrate it cleanly?

Or would it just copy chaos faster?

Data sovereignty is not a buzzword.
It is becoming a strategic defensive position.

We are not currently experiencing any tool hype.

We are witnessing a reassessment of power relations in the software market.

And that's far more exciting than any "SaaS is dead" headline.