Anthropic's internal effort to have Claude run growth experiments end to end.
From Amol Avasare, Anthropic head of growth
CASH breaks growth shipping into four stages Claude is scored on. Anthropic's CASH effort splits the shipping loop into finding opportunities, building, testing against quality and brand bars, and analyzing results. Each stage is evaluated and scored on how well Claude does it, then hill climbed weekly.
Brand guidelines can be encoded as a skill so the testing stage needs less human review over time. Amol Avasare, Anthropic's head of growth, frames it as part of a broader push to have AI run routine growth work so the team can focus on large bets.
Start automation on copy and small UI tweaks, with a human approving. CASH currently handles copy changes and minor UI tweaks, with a human approving each change. Its win rate is roughly that of a PM two to three years into the job.
It showed no results before Opus 4.5 and became promising with Opus 4.6. Amol says "it ultimately prints money," but the win rate comes from him alone, with no data shown.
Claude already does routine growth monitoring for Anthropic's head of growth. Amol has Claude scan 20 to 25 charts each morning, flag misalignment in Slack weekly, and give him manager-style feedback.
He puts Claude's current growth and PM work at about a junior PM's level, while cross-functional alignment stays a human job.
CASH splits shipping into finding opportunities, building, testing and analyzing, and each step is scored and improved weekly. It handles copy changes and minor UI tweaks with a human approving, at about the win rate of a PM two to three years in; it showed no results before Opus 4.5.
CASH breaks growth shipping into four stages Claude is scored on. Anthropic's CASH effort splits the shipping loop into finding opportunities, building, testing against quality and brand bars, and analyzing results. Each stage is evaluated and scored on how well Claude does it, then hill climbed weekly.
Brand guidelines can be encoded as a skill so the testing stage needs less human review over time. Amol Avasare, Anthropic's head of growth, frames it as part of a broader push to have AI run routine growth work so the team can focus on large bets.
Start automation on copy and small UI tweaks, with a human approving. CASH currently handles copy changes and minor UI tweaks, with a human approving each change. Its win rate is roughly that of a PM two to three years into the job.
It showed no results before Opus 4.5 and became promising with Opus 4.6. Amol says "it ultimately prints money," but the win rate comes from him alone, with no data shown.
Claude already does routine growth monitoring for Anthropic's head of growth. Amol has Claude scan 20 to 25 charts each morning, flag misalignment in Slack weekly, and give him manager-style feedback.
He puts Claude's current growth and PM work at about a junior PM's level, while cross-functional alignment stays a human job.
Developer toolsA chip company whose GPUs and CUDA software power gaming, AI and accelerated computing.
Developer toolsAnthropic's agentic coding tool that lets Claude work on code end to end.
Developer toolsAn AI code editor that lets engineers build software with AI while staying in control.
Developer toolsA code hosting and collaboration platform for software developers, with free and paid plans.
Developer toolsAn AI app builder that lets people create software by chatting with an agent.
An app development company that builds apps for founders and clients.
Developer toolsA fast, well-designed project and issue tracking tool for software teams.
Developer toolsGoogle's backend platform for building apps.