For the last 20 years, SaaS has meant Software as a Service.

Log in. Use the tool. People do the work.

That model is bending.

The next version of SaaS is Synthetic as a Service.

Old SaaS gave you dashboards.

Synthetic as a Service gives you teammates.

Old SaaS sold seats.

The next generation sells outcomes.

Old SaaS sold seats. The next generation sells outcomes.
From software access to synthetic capacity.

We are already seeing the transition in software development.

Increasingly, strong software engineers aren’t spending their entire day manually typing code. They are directing synthetic systems, reviewing their output, designing architectures, debugging edge cases, and deciding what should be built.

They are moving up the stack.

Moving up the stack

As synthetic intelligence performs more execution, humans move toward judgment, architecture, and direction.

That same transition will increasingly happen across knowledge work.

Instead of buying a CRM, you may get a synthetic sales teammate that researches accounts, drafts outreach, updates the CRM, follows up, and books meetings.

Instead of buying project-management software, you may get a synthetic operations coordinator that watches projects, identifies blockers, follows up with people and systems, and produces the weekly brief.

Instead of buying another analytics dashboard, you may get a synthetic analyst that investigates the data and tells you what changed, why it matters, and what should happen next.

The software doesn’t disappear.

But increasingly, you stop operating it.

The synthetic worker operates it for you.

That changes the economics.

Historically, companies bought human capacity in a few familiar ways.

You hired an employee and paid for their time.

You hired a consultant and paid hourly.

Or you hired a consultant for a defined outcome or project.

Synthetic workers will likely follow similar models.

Some synthetic capacity may effectively be metered — potentially through tokens, compute, tasks, or some other unit of consumed intelligence.

Other synthetic services will simply be priced around the outcome.

Resolved ticket.

Qualified lead.

Completed analysis.

Managed workflow.

Delivered project.

New labor economics

Companies have always purchased capacity. The substrate is changing.

That makes traditional per-seat pricing increasingly awkward.

Seats made sense when humans were the operators.

Synthetic workers don’t map naturally to seats.

They map to work.

And that raises a much more important question than whether the underlying model is intelligent:

Can you trust it?

The trust layer

Once intelligence can act, trust becomes infrastructure.

A chat window isn’t a teammate.

A teammate needs appropriate permissions.

It needs controlled access to systems and data.

It needs security boundaries.

It needs memory boundaries.

It needs auditability.

It needs to understand when it can act autonomously and when a human must approve the next step.

And it has to reliably follow through.

Once synthetic intelligence can actually take actions inside a business, security and access control stop being secondary IT concerns.

They become part of the product itself.

The winners in this next generation won’t simply wrap an intelligent model in a nice interface.

They’ll build trusted systems around intelligence — workflows, tools, permissions, security, memory, audit trails, domain knowledge, and human oversight.

That is also the direction I am now pursuing through Stoin.

We aren’t particularly interested in adding another dashboard to the world.

We’re interested in helping companies turn synthetic intelligence into actual operating capacity: systems that can research, coordinate, analyze, execute workflows, interact with existing software, and increasingly complete meaningful work alongside humans.

Humans don’t disappear in that model.

They move up the stack.

Less clicking.

Less copying.

Less chasing.

More deciding.

More designing.

More creating.

More leading.

For two decades, the defining software business model was:

Software as a Service.

The next one may have almost the same acronym.

But a very different meaning.

Synthetic as a Service.

We won’t rent software forever. We’ll hire synthetic capacity.