Layer 2 — AI Agents · Chapter 6
Business Opportunities and Outlook
Agents may change software economics because customers will increasingly pay for completed work rather than seats, clicks, or access to a tool.
Strong early opportunities
- Back-office coordination across email, documents, and enterprise systems
- Customer-service resolution rather than simple question answering
- Research and monitoring that require repeated checking
- Software maintenance and testing
- Scheduling, procurement, logistics, and vendor management
- Compliance evidence collection and recurring reporting
Business models
Agent companies may charge per completed task, per resolved case, by transaction volume, through subscriptions, or as a managed service. Outcome-based pricing can align value with customer benefit, but it also makes reliability more important.
Defensibility
Durable advantages may come from proprietary workflow data, deep integrations, trusted permissions, domain expertise, customer distribution, and superior evaluation systems.
Likely progression
The near-term market will favor bounded agents that operate inside clearly defined workflows. Over time, the boundary may expand as systems improve at memory, planning, verification, and long-running tasks.
Connection to Layer 1
Agents do not replace the application layer. They change it. Many future applications will become interfaces through which users assign goals, supervise work, and review outcomes.