Amol Avasare on Growth at Anthropic, AI-Native PM Work, and Freedom Through Constraints

Amol Avasare on Growth at Anthropic, AI-Native PM Work, and Freedom Through Constraints

transcriptlennys-podcastgrowthanthropicaiproduct-managementactivationculture

Amol Avasare on Growth at Anthropic, AI-Native PM Work, and Freedom Through Constraints

Source: Lenny’s Podcast Speaker: Amol Avasare Date: ~2025 Link: https://www.lennysnewsletter.com/p/anthropics-1b-to-19b-growth-run

Key ideas

  • “Success disasters” dominate the activation problem. Roughly 70% of Amol’s growth work is spent on users who have an early win but churn before embedding Claude into their workflows — success disasters. Activation is not about getting users to any first interaction; it is about guiding them past good friction (engaging, meaningful use) and eliminating bad friction (unnecessary barriers). The ChatGPT memory import feature is a tactical example: when a competitor launched an advantage, Anthropic reduced the switching cost for users migrating in, turning a threat into an activation lever.
  • CASH: Claude growing Claude. CASH (Claude Accelerates Sustainable Hypergrowth) is an automated growth-experimentation system where Claude generates A/B test hypotheses, runs them, analyses results, and proposes the next round — removing humans from most of the experimentation loop. The growth team’s role shifts from ideating tests to reviewing outputs and setting direction. As Amol put it, Claude is now effectively growing itself.
  • AI-native PM practices at scale. The growth team runs scheduled Claude/Cowork tasks every morning: 20–25 charts reviewed autonomously, with anomalies surfaced to Amol before he opens his laptop. Weekly alignment scans cross-reference Slack, project goals, and OKRs to surface cross-functional misalignment before it causes spin. Amol even asks Claude to simulate his manager Ami Vora‘s feedback on his own week, drawing on her public writing and internal Slack. Admin (expenses, reimbursements, meeting bookings) is fully delegated to Claude. This is less AI-assisted work and more AI-orchestrated work with human review.
  • Leaving money on the table is a growth principle, not a weakness. Growth teams that try to squeeze every last metric point drift toward dark patterns and brand erosion. Amol operates a two-category filter for controversial tests: Category 1 — red lines where the test simply will not ship regardless of results (AI safety, brand, values); Category 2 — elevated sensitivity but runnable if the hypothesis is strong and the upside justifies the discomfort. Anthropic’s PBC (public benefit corporation) structure legally embeds this trade-off: maximising shareholder value is not the overarching goal.
  • Focus as competitive advantage; freedom through constraints. Anthropic’s early, narrow focus on B2B and coding was driven by necessity — least-funded player, no first-mover advantage, no distribution. A Ben Mann document from 2021 (pre-ChatGPT) already argued for going deep on AI coding. The principle generalises: constraints eliminate excess choice and clarify the path. Amol applies it personally too — a traumatic brain injury from MMA in early 2022 forced nine months off work and a complete rebuild of his daily habits (no alcohol, no caffeine, mandatory breaks, annual meditation retreats). The habits that were imposed by injury became durable competitive advantages.

Overview

Amol Avasare — Head of Growth at Anthropic, formerly at Mercury and Masterclass — covers Anthropic’s $1B to $19B ARR run in 14 months, the activation-first framing behind that growth, and how the ~40-person growth org is itself an AI-native operation. The episode moves from growth tactics (success disasters, activation, CASH) through culture (talent density, Anthropic’s mission-first DNA, notebook Slack channels) to personal resilience (TBI recovery, freedom through constraints). One of the more technically and personally candid growth episodes on Lenny’s Podcast.