Helium

Helium charges customers based on the incremental revenue their AI generates through continuous experimentation on user funnels. ​​‌‌‌‌‌‌‌​‌‌​​‌​​​​​​‌‌​‌‌‌​​​‌‌Rather than a flat SaaS fee, they structured pricing as a percentage of uplift—aligning incentives directly with customer outcomes. To validate willingness to pay, they ran their system with their first customer, a social app with 20M monthly active users, testing new in-app paywalls. The customer saw measurable subscription lift within weeks, providing concrete proof that the AI's experiments delivered real revenue gains. This outcome-based model meant Helium only earned when customers succeeded, eliminating pricing objections. The early signal that validated their approach was immediate: the customer renewed and expanded usage without negotiation, demonstrating they'd happily pay for proven incremental revenue. The team's pedigree—Zach's $100M+ growth experiments at Uber, Shishir's LLM work at Alexa, and Anish's Meta ML expertise—gave them credibility to execute this complex technical solution, but the customer's rapid expansion proved the market genuinely valued self-improving software.