What happens to SaaS revenue when AI works exactly as promised?

For two decades, enterprise software benefited from a remarkably simple commercial equation:

Revenue = Seats × Price per Seat

Agentic software is beginning to break that relationship.

AI can allow customers to accomplish more work with fewer human operators. The product becomes more capable, the customer receives more value—and a SaaS company that still charges exclusively by human seat can end up earning less.

AFTER THE SEAT examines the emerging migration of the software revenue denominator:

Seat → Usage → Action → Outcome → Permission → Settlement

Rather than arguing that per-seat SaaS is disappearing, Haijian Xing develops a practical framework for identifying where different software businesses may move next.

The book introduces:

the Seat Exposure Audit, for identifying ARR vulnerable to AI-driven headcount compression;
the AI Revenue Replacement Ratio, for measuring whether new machine revenue is replacing seat revenue lost to automation;
the SaaS Strategic Migration Matrix: Defend the Seat, Protect With Usage, Replace the Denominator, or Become the Tollbooth;
the AI Economic Event, a commercial unit connecting identity, purpose, usage, permission, cost, and outcome;
the distinction between internal infrastructure meters and external customer-value units;
permission, revocability, rating, and settlement as emerging dimensions of agentic software;
and the Protocol Tollbooth, where strategic value comes from governing economically important machine passage rather than merely owning an application.

The central question is no longer simply:

How should we charge for AI?

It is:

What economic event will our company own when the human seat is no longer the only meaningful unit of software consumption?

This Public Framework Edition is designed to make the analytical framework widely accessible. It identifies the architecture and the questions. Applying that architecture to a specific company—its ARR, products, workflows, permissions, economic events, and migration path—is a separate strategic task.

The framework is meant to spread.
The implementation is where the work begins.

That last distinction is important because the original research explicitly moves from applications as destinations toward systems that control economically necessary routes.