Lenny's PodcastAmol Avasare: Why 70% of Growth Work Is Firefighting Wins
Through 'success disaster' firefighting and capability-overhang activation; Anthropic's growth team turns hypergrowth chaos into compounding wins.
At a glance
WHAT IT’S REALLY ABOUT
Anthropic growth leader on activation, automation, safety, and future roles
- Anthropic’s growth org looks traditional on paper (acquisition, activation, monetization) but Amol says most effort goes to “success disasters,” where rapid growth breaks systems that must be urgently repaired.
- AI product activation is uniquely hard because capabilities evolve faster than onboarding and users often don’t know what to ask for, so Anthropic uses targeted onboarding questions and “good friction” to route users to the right value.
- Anthropic is actively automating growth experimentation via CASH (Claude Accelerates Sustainable Hypergrowth), with Claude generating, building, and analyzing small experiments at roughly a junior-PM win rate—improving quickly as models advance.
- As engineering productivity jumps most from AI tooling (e.g., Claude Code), PM and design become the bottleneck, pushing teams toward more PMs and/or “mini-PM” engineers for small projects.
- Anthropic’s focus on coding and B2B, paired with a safety-first mission (including PBC structure) and a highly transparent “notebook channel” culture, is presented as a core competitive advantage rather than a constraint.
IDEAS WORTH REMEMBERING
5 ideasIn hypergrowth, growth work becomes firefighting of “success disasters.”
Amol estimates ~70% of his time is spent on problems caused by things going too well (capacity, UX scaling issues, monetization edge cases), even while metrics look uniformly green.
Activation is the biggest AI-product bottleneck because users under-ask and models over-improve.
Users often default to trivial queries, while model capabilities shift faster than product teams can translate them into onboarding and guided workflows, making “capability diffusion” a core product challenge.
Add friction when it helps match the right user to the right value.
Anthropic and Amol’s prior work (Mercury, Masterclass) show that extra steps—questions, quizzes, split forms—can increase completion by reducing cognitive load and improving personalization, as long as it’s value-adding rather than annoying.
AI-first businesses should bias toward larger growth bets, not just micro-optimizations.
Amol argues AI-first product value can be 100–1000x higher in a couple years, so teams should avoid “missing the forest” by over-indexing on small wins—even if 1% improvements are massive at scale.
Automated experimentation is arriving first in growth because loops are measurable and repeatable.
CASH operationalizes an experiment loop (opportunity → build → QA/brand → analysis), already producing wins on copy/UI tweaks with humans approving outputs today and less review needed as brand/safety constraints become machine-checkable.
WORDS WORTH SAVING
5 quotesHistorically, we were very much the smallest, least well-funded player in this space. ... It’s a complete miracle that we’ve gotten to the stage that we have.
— Amol Avasare
Roughly 70% of what I spend my time on is what we internally refer to as success disasters.
— Amol Avasare
Activation is a really big challenge in AI.
— Amol Avasare
Adding friction and adding the right steps leads to higher conversion... Just cut all the steps and get them into the product—like that doesn’t work most times.
— Amol Avasare
We are starting to look at how do we automate growth... CASH... Claude Accelerates Sustainable Hypergrowth.
— Amol Avasare
High quality AI-generated summary created from speaker-labeled transcript.
Get more out of YouTube videos.
High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.
Add to Chrome