Long-running AI agents and the case for outcome-based pricing

The essay argues that agents able to maintain context and work across weeks or months could manage extended customer journeys, such as plaintiff law case preparation. It suggests vendors could charge based on work products or outcomes, rather than focus only on cost savings. The author outlines three approaches: selling software, taking over an outsourced service, or operating an AI-native business. Legal-fee restrictions and disputes over credit for results could complicate outcome-based pricing.