we unpack one of the fastest built experimentation engines i've seen!! In this episode, I sit down with Kameron Tanseli (Growth Engineering) and Rosie Hoggmascall (Growth & Product) from Fyxer.ai and we retrace the steps of building a 500+ /year experimentation program.

We go deep into:

• Why their first onboarding experiments actually made things worse
• The tooling shift from PostHog to GrowthBook (and why it mattered)
• How decentralized ownership (no PM, no heavy process) unlocked serious speed
• Why every engineer acts like a founder
• How Slack transparency builds experimentation culture
• Using AI agents to automate experiment reporting, analysis, and cleanup
• What “time to value” really means in an AI product
• The emotional side of onboarding (“inbox shock”) and how it impacts retention
• How to prioritize when you’re scaling from 5 → 70+ people

We also explore:

• ➤ The difference between growth and product prioritization
• ➤ How to surface insights across support, sales, and product
• ➤ What’s still impossible to automate in experimentation
• ➤ The tracking challenges of building inside Gmail and Outlook and of other AI tools