13 Business Models Worth Building in the AI Agent Era

Greg Isenberg outlines 13 business categories he sees as viable in a world of AI agents and abundant automation. The framework prioritizes owning assets or audiences that retain value as agents handle more work: AI-native service firms where humans review exceptions, offline businesses selling physical spaces and labor, proprietary data sets, domain-specific tools, robotics, physical products with loyal followings, compute and energy infrastructure, health and longevity services, marketplaces and social networks, real assets, vertical agents for specific industries, and security.

Isenberg argues that in an agent-driven economy, scarce inputs like compute, energy, and security become critical, while organic audiences matter more than paid ads as ad costs climb. He flags proprietary data and domain-specific harnesses as the "two gold rushes" currently unfolding, and notes that businesses selling rooms, objects, and labor that agents only support—not replace—will endure.