Eric and Neil break down how Eric cut his AI token spend from around $7,500 a month to nearly $0 by changing his model hierarchy, fixing fallback issues, and reducing unnecessary API usage. They also get into why usage-based AI pricing is changing software, why some tools become more valuable in an agent-driven world, and what founders, marketers, and agencies need to understand as AI costs shift from seat-based pricing to usage-based pricing. Key takeaways ◾ You can dramatically reduce AI token spend by fixing model hierarchy and fallback logic. ◾ AI costs need to be actively monitored because broken workflows can quietly burn cash. ◾ Usage-based pricing is becoming a bigger part of software economics. ◾ Some tools get more valuable in an agent-first world, while others matter less. ◾ Agencies that help companies become AI-readable may have a major opportunity. Chapters (00:00) How Eric cut his $7,500 AI token spend (03:25) Why usage-based AI pricing is going up (05:08) Why some software matters less in an agent world (08:11) ClickFlow ad break (12:29) Why AI-readable brands matter more (17:19) What this means for agencies and founders