Automation Anywhere's acquisition of Aisera this week includes a striking claim:

That's not just a pricing change. It's a bet that the infrastructure to actually charge for outcomes, not seats, finally exists. And if you understand what that infrastructure requires, that's a bold claim.

The Infrastructure Gap Between Promise and Proof

Outcome-based pricing isn't new. Vendors have been promising it for years. What's been missing is the infrastructure to actually deliver it.

You need telemetry that proves what happened. Observability that catches failures before customers do. Governance that handles authentication breaks and workflow changes gracefully. Most vendors claiming "outcome-based pricing" are just doing seat-based pricing with creative accounting.

At TinyFish, where we build enterprise web agent infrastructure, the pattern is consistent: the gap between "our agent can do X" (demo) and "we can charge you for X because we can prove it happened reliably" (production) is enormous. That's what makes this acquisition interesting. It's either the first real attempt to build that infrastructure, or it's expensive theater.

What Aisera brings: integration with 100+ enterprise applications and 3,000+ built-in workflows across ServiceNow, Jira, Salesforce, Workday. That's not surface-level automation. That's deep integration handling authentication across multi-tenant environments, managing state across systems, orchestrating outcomes where failures cascade. When you're charging for outcomes instead of seats, that integration depth becomes critical infrastructure. You need to measure what actually happened and prove the agent resolved the ticket correctly.

How Lock-In Changes When Agents Do the Work

This acquisition reveals something beyond just combining two companies: it shows how switching costs evolve when agents replace seats.

Traditional ITSM lock-in comes from implementation complexity. ServiceNow implementations average 1,200-1,400 hours over months, with costs 60% higher than alternatives. But when agents are doing the work, the switching costs shift to operational dependencies. If your incident resolution depends on how agents handle authentication patterns, route tickets across your specific stack of enterprise systems, and orchestrate outcomes reliably, switching means rebuilding those operational processes.

In web agent infrastructure, customers build workflows around agent outputs. The lock-in isn't the seat count. It's the operational processes that depend on how agents navigate your specific environment. When Aisera's platform learns your authentication patterns, integrates with your particular mix of 100+ enterprise apps, and handles your workflow variations, that creates switching costs that replace traditional seat-based lock-in.

Agents reducing seats is already happening. Research shows 40% of enterprise buyers now cite seat reduction as their primary cost lever. What matters is whether the infrastructure that makes agents reliable at scale creates new forms of value capture.

Three Signals That Would Prove This Works

If this acquisition works—if Automation Anywhere can actually charge for outcomes instead of seats at enterprise scale—watch for three signals:

• Trusted telemetry: Can they prove what agents accomplished with telemetry customers trust?
• Graceful failure handling: Do they handle authentication and workflow failures gracefully enough that customers depend on them for operational processes?
• Market validation: Do other enterprise software vendors follow, or does this remain an isolated experiment?

The infrastructure to measure and charge for outcomes reliably, not just promise to, will define what software economics look like when agents replace seats. This acquisition is the first major test of whether that infrastructure actually exists.