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.
CHAPTERS
- 0:00 – 3:54
Anthropic’s unprecedented ARR curve and the realities behind “fastest-growing”
Lenny sets the stage with Anthropic’s extraordinary revenue trajectory and asks what it’s like to lead growth in the middle of that kind of acceleration. Amol frames the company as research- and model-led first, with growth riding a wave of rapidly compounding capabilities.
- •ARR growth from ~$1B to ~$19B+ in ~14 months and why linear charts stop being useful
- •Anthropic’s historical disadvantages (less funding, less distribution, no first-mover)
- •Growth as a company-wide outcome, not just a growth-team output
- •Operating in “log-linear” mode as the business scales exponentially
- 3:54 – 8:36
How Amol cold-emailed Mike Krieger into a job that didn’t exist yet
Amol explains how he noticed Claude lacked a dedicated growth function and reached out directly to Anthropic’s CPO. He breaks down the mechanics of cold email: subject lines, channel selection, brevity, and persistent follow-up.
- •Anthropic wasn’t hiring for growth roles when Amol reached out
- •Cold email fundamentals: optimize opens, keep message short, be clear about fit
- •Why personal email and avoiding crowded channels can matter
- •Follow-up discipline: persistence until you get a definitive “no”
- 8:36 – 11:21
What it’s like to run growth inside a rocket ship: “success disasters”
Amol describes the emotional and operational intensity of scaling at Anthropic. Much of his time goes to “success disasters”—situations where things go so well that systems, processes, and product surfaces break under demand.
- •Growth team impact is real, but model quality and research are the primary drivers
- •The planning whiplash of aggressive targets becoming reality
- •“Success disasters” as the dominant workload during hypergrowth
- •Balancing firefighting with proactive, strategic growth work
- 11:21 – 15:19
What Anthropic’s growth team actually does (and why activation dominates)
Amol outlines classic growth pillars—acquisition, activation, and monetization—while emphasizing that AI products amplify activation challenges. They’re constantly working to reduce cold-start pain and quickly connect users to the right workflows.
- •Growth scope: acquisition, activation, monetization, pricing/packaging
- •AI-specific activation: getting users to meaningful early value is harder than in traditional products
- •“Importing memory” as a cold-start tactic tied to user context and personalization
- •The goal: faster path to ‘Claude understands me’ and ‘Claude helps me’
- 15:19 – 18:05
Why activation is uniquely hard in AI: capability overhang and shifting priors
Amol explains that models improve so quickly that onboarding learnings can become obsolete before teams can operationalize them. The key is guiding users to high-leverage use cases and using the right kind of friction to match users to the right product surfaces.
- •Activation’s outsized downstream effect on retention is even stronger for AI products
- •Capability overhang: model power outpaces user behavior and product education
- •The challenge of running tests when model releases continuously change what’s possible
- •Core tactic: identify user characteristics and route them to the right features—even if it adds steps
- 18:05 – 25:10
Mercury onboarding turnaround: quality as a growth lever (and friction that helps)
Amol shares a Mercury case study where the growth team spent a quarter focused on onboarding quality rather than incremental metric chasing—leading to major conversion gains. He generalizes the lesson: remove annoying friction, but keep (or add) friction that improves understanding and relevance.
- •Regulated onboarding is complex; first impressions matter disproportionately
- •“Forget metrics, improve quality” led to one of Amol’s most impactful quarters
- •Good friction reduces cognitive load and increases completion (e.g., splitting forms, guided quizzes)
- •Friction works when it helps users feel the product is ‘for them’ and clarifies next steps
- 25:10 – 26:17
Inside Anthropic Growth: pods, horizontals, and cross-functional makeup
Amol describes a ~40-person growth org with platform/monetization horizontals and audience- or product-aligned pods. The structure is designed for focus across multiple products and for tight coupling with key product teams.
- •Org size and composition: PMs, engineers, designers, data
- •Horizontals: growth platform and monetization across products
- •Audience/product pods: B2B, Claude Code, knowledge workers, API, etc.
- •Org design goal: focus + strong cross-functional tie-ins in a multi-product company
- 26:17 – 33:47
Why Anthropic prioritizes big swings over micro-optimizations: growth in an exponential world
Despite the massive dollar impact of small percentage wins at Anthropic’s scale, the team intentionally leans toward bigger bets. Amol argues that when AI is the core value engine, product value can increase by orders of magnitude—so missing the next market matters more than squeezing the current funnel.
- •Traditional growth: mostly small/medium experiments; Anthropic shifts toward larger bets
- •Exponential framing: AI-first product value in 1–2 years could be 100–1000x
- •New markets appear quickly (e.g., agentic coding) and can dwarf prior opportunities
- •Guidance: if AI underpins the core value prop, bias toward bigger, producty bets
- 33:47 – 38:14
Automating growth experimentation with Claude (CASH): “press play” growth loops
Amol introduces CASH (Claude Accelerates Sustainable Hypergrowth), an initiative to automate growth experimentation end-to-end. They evaluate model performance across a loop: identify opportunities, build, QA/brand check, and analyze results—already seeing measurable wins on smaller changes.
- •CASH goal: use Claude to automate growth experiments, starting with small UI/copy changes
- •Experiment loop breakdown: opportunity discovery → build → quality/brand → analysis
- •Current performance: comparable to a junior PM’s win rate, improving rapidly with new models
- •Human-in-the-loop review today, with a path to decreasing oversight via brand ‘skills’/guardrails
- 38:14 – 41:07
AI that suggests what to build: where it works today—and what still blocks it
The conversation shifts from AI executing tasks to AI proposing the right work to do next. Amol highlights that stakeholder alignment remains a major constraint for big bets, even as AI improves at drafting, testing, and analysis.
- •AI moving “from the middle outward”: not just building, but also recommending experiments
- •Why growth is an early wedge: data-driven loops with frequent iteration
- •Brand and governance constraints can be encoded, reducing the need for manual review
- •Cross-functional stakeholder management remains the hardest-to-automate bottleneck
- 41:07 – 56:06
The future of PM/engineering/design: ratios, mini-PMs, and the decline of PRDs
Amol argues engineering is currently getting the biggest productivity lift from AI tools, which strains PM and design capacity. Anthropic responds by hiring more PMs in some areas and deputizing product-minded engineers as “mini PMs” for small projects, while reducing reliance on heavy documentation.
- •Engineering leverage is compounding faster than PM/design leverage (today)
- •Potential future: more PMs to keep up with accelerated engineering throughput
- •Rule of thumb: <2 engineering weeks → engineer drives as mini-PM; larger work → PM accountable
- •PRDs increasingly optional; kickoffs, prototypes, and fast alignment replace heavy docs
- 56:06 – 1:06:27
How Amol uses AI daily: charts, misalignment detection, admin automation, and AI coaching
Amol shares concrete workflows where Cowork/Claude proactively summarizes metrics, flags org misalignment from Slack context, and automates routine admin work. He also describes using an “AI version” of his manager to generate weekly feedback and coaching prompts.
- •Scheduled daily/weekly agents: scan 20–25 charts, highlight anomalies and insights
- •Slack MCP-based workflow to detect potential misalignment across teams
- •Automating “life admin”: rooms, inbox triage, reimbursements/expenses
- •Manager/IC coaching: generate feedback for directs; simulate manager feedback for self-improvement
- 1:06:27 – 1:12:07
Why Anthropic won: focus on coding + B2B, leadership clarity, and constraint-driven strategy
Amol attributes much of Anthropic’s success to early strategic focus and strong leadership, including a long-running bet on coding. He explains how constraints forced sharper choices, and how open internal communication (including “notebook channels”) helps scale alignment and culture.
- •Early strategic clarity: deep focus on coding and B2B before it was obvious
- •Coding as a research accelerant: better tools → faster research loop → stronger models
- •Constraints as an advantage: less funding/distribution forced focus to reach escape velocity
- •Culture mechanics: notebook channels, openness to disagreement, and scaling shared beliefs
- 1:12:07 – 1:35:12
Balancing hypergrowth with AI safety—and being willing to leave money on the table
Amol explains how safety is structurally and culturally embedded at Anthropic, influencing which growth tactics are acceptable. He also describes a broader growth principle: protect brand, user trust, and long-term outcomes rather than squeezing every short-term metric.
- •Anthropic’s mission: ensure powerful AI benefits humanity; safety is a first-order constraint
- •Corporate structure: Public Benefit Corporation enables non-shareholder-maximizing tradeoffs
- •Decision framework: some tests are “never run”; others require high return to justify any ‘ick’
- •Long-term view: safety posture and trust can become a durable competitive advantage
- 1:35:12 – 1:52:48
Failure corner and personal resilience: startup shutdown, traumatic brain injury, and rebuilding
Amol recounts shutting down a funded mental health startup and the hard-earned lessons about transparency and long timelines. He then shares the life-altering traumatic brain injury that forced months of recovery, reshaped his habits, and deepened his commitment to meditation and sustainable performance.
- •Startup failure: shutting down after raising money; the importance of honest investor updates
- •TBI recovery: months unable to work normally; slow exposure-based rebuilding; reinjury risks
- •New operating system: no alcohol/caffeine, mandatory breaks, annual meditation retreats
- •Mindset shift: acceptance, “freedom through constraints,” and contentment without outcomes
