What to expose to general agents vs. what to keep in-house
The author argues that as general AI agents aggregate software features, SaaS companies must strategically choose whether to expose low-stakes tasks to third-party agents or invest in proprietary, high-stakes jobs where they can maintain accountability and outcomes. Using examples like HubSpot and Uber, the analysis suggests that survival depends on building an integrated 'package of emphasis' that general labs cannot easily replicate. Open article →