Reverse SaaS

Welcome Back to XcessAI

For the past two decades, the dominant technology model was simple.

Build software.
Replace services.
Scale infinitely.

Software ate services, in what became known as SaaS (Software as a Service).

But AI may be reversing part of that equation.

The latest generation of frontier AI models has dramatically expanded what software systems can do on behalf of users. Tasks that once required navigating multiple applications can increasingly be executed through simple instructions.

AI-assisted coding has lowered the barrier to building software dramatically. Capabilities that once required large engineering teams can now be prototyped by small teams or even individuals.

The result is visible both in the rapid proliferation of AI-built applications and in the growing debate about the long-term defensibility of traditional SaaS models.

Some have started calling this shift Service as Software - hence “Reverse SaaS.”

We believe this is a very big event, so let’s dive a bit deeper to understand the potential implications for business.

The application layer problem

A large share of modern software does not exist to compute.

It exists to organise work.

Forms.
Dashboards.
Workflow interfaces.
Database navigation.

Much of the modern application layer is an interface for humans to coordinate information.

AI collapses that interface.

Instead of navigating software, users increasingly instruct systems.

The interaction moves from interface navigation to goal specification.

From clicking.

To delegating.

What happens when the interface collapses

When software interfaces disappear, something interesting happens.

The value of software shifts.

Historically, value lived in:

• the interface

• the workflow

• the application layer

But when AI can directly operate on systems, those layers become thinner.

The advantage moves elsewhere.

It moves to expertise.

The rise of service as software

When AI systems can execute complex tasks, expertise becomes programmable.

Not perfectly, but enough to scale. For example:

• A marketing agency can encode its campaign playbooks into AI workflows.

• A legal practice can encode research frameworks and document logic.

• A finance team can encode operating models and decision frameworks.

The result is something new.

Services delivered with software-like scalability.

Historically, services did not scale because expertise was locked inside people. AI begins to externalise that expertise into systems. When playbooks, judgment, and decision frameworks are encoded into workflows, the economics change. The marginal cost of delivering expertise begins to resemble software.

A hybrid model: service expertise delivered with software-like scalability.

Consider finance. Traditionally, access to high-quality strategic finance advice required hiring experienced CFOs or engaging expensive advisory firms. AI-assisted systems can now encode operating models, capital allocation frameworks, and financial diagnostics into scalable tools. The expertise remains human, but its delivery becomes software-like.

Why expertise becomes the moat

AI models are improving rapidly.

But models alone do not know what “good” looks like in a specific domain.

That knowledge lives in practitioners.

In playbooks.
Judgment.
Pattern recognition.
Institutional experience.

When that expertise is encoded into AI-assisted systems, it becomes scalable.

The advantage shifts from:

who built the software

to

who understands the problem deeply enough to structure the system.

The economic implication

When intelligence becomes abundant, structure becomes scarce.

Software platforms provide the tools.

But the value increasingly sits in:

• domain expertise

• structured playbooks

• proprietary context

• execution frameworks

In other words:

The service layer.

What this means for builders

For many builders, the opportunity may not be to build another application.

It may be to encode expertise.

To turn professional knowledge into scalable delivery systems.

Instead of selling software.

They sell outcomes.

Software becomes the delivery mechanism.

Expertise becomes the product.

Closing

Software once replaced services.

AI may be turning services into software.

The companies that win may not be the ones building the most interfaces.

They may be the ones encoding the deepest expertise.

Until next time,
Stay adaptive. Stay strategic.
And keep exploring the frontier of AI.

Fabio Lopes
XcessAI

💡Next week: I’m breaking down one of the most misunderstood AI shifts happening right now. Stay tuned. Subscribe above.

Read our previous episodes online!