ExpertsFMHead of Growth (Anthropic): “Claude is growing itself at this point” | Amol Avasare
Lenny's Podcast: Product | Career | Growth · April 5, 2026
5 Selected Lessons:
case · 2 heuristic · 1 reasoning_pattern · 1 priority · 1
Safety and mission over commercial growth, formalized structurally
Source
Lenny's Podcast: Product | Career | Growth — Head of Growth (Anthropic): “Claude is growing itself at this point” | Amol Avasare (April 5, 2026) · at 1:12:09 ▸
When this applies
when an organization must choose between a near-term growth/revenue opportunity and a stated mission or ethical constraint
The takeaway
If a value must consistently beat growth in trade-off decisions, encode it structurally (legal form, explicit decision frameworks) rather than relying on case-by-case willpower.
Why it works
Encoding the priority into corporate structure (public benefit corporation vs. Delaware C-Corp) removes the fiduciary obligation to maximize shareholder value as the sole objective, making safety-first decisions legally and organizationally defensible rather than just aspirational.
One level up
Stated values only reliably win trade-offs against growth when they are built into the decision architecture, not left as sentiment.
Where it stops holding
This stops working if the structural commitment isn't backed by genuine leadership conviction — a PBC structure without real belief would just be cosmetic, per Amol's emphasis that leadership 'really believe[s] in this a lot.'
Detail
hierarchy:
- 1. AI safety / 'this needs to go well for humanity' — non-negotiable, company exists for this
- 2. Brand, customer friendliness, and values — high threshold required to override
- 3. Growth metrics / commercial upside — explicitly subordinate to the above
rationale: Anthropic delayed launching Claude before ChatGPT for safety reasons, taking a major first-mover hit; the company nonetheless became the fastest-growing in history, which Amol reads as evidence the trade-off paid off long-term.
In their words
created a public benefit corporation, PBC, which allows you to legally say that maximizing shareholder value is not the overarching umbrella goal of this company, and you can optimize for public benefit.
Cross-functional alignment is the last mile AI won't automate soon
Source
Lenny's Podcast: Product | Career | Growth — Head of Growth (Anthropic): “Claude is growing itself at this point” | Amol Avasare (April 5, 2026) · at 37:22 ▸ · at 40:06 ▸
When this applies
when deciding which parts of any coordination-heavy role can be automated vs. which require human judgment
The takeaway
Expect automation to eat the structured, evaluable stages of a workflow first and leave the messy human-negotiation layer for last.
Why it works
Human stakeholder alignment requires navigating competing incentives, politics, and trust that current AI agents can't fully substitute for, even as AI automates the mechanical parts of idea generation, building, and testing.
One level up
The hardest problem in any human system is alignment among humans, not raw task execution — and that bottleneck persists even as capability elsewhere compounds.
Where it stops holding
This stops working, per the speakers' own joke, only once other stakeholders are themselves represented by autonomous agents negotiating with each other.
Detail
approach: List out the stages of a workflow that AI is automating (ideation, build, test, analysis) and note the one stage (cross-functional stakeholder management) that still needs human brains.
signals:
- disagreement among stakeholders with strong opinions
- need for negotiation/mediation
- 6-person-room alignment problems
rationale: Joel's quip: 'we will have AGI, and it will still be impossible to get six people in a room to align.'
In their words
he pings me, he's just like, Amol, we will have AGI, and it will still be impossible to get six people in a room to get to align.
Good friction vs bad friction
Source
Lenny's Podcast: Product | Career | Growth — Head of Growth (Anthropic): “Claude is growing itself at this point” | Amol Avasare (April 5, 2026) · at 13:55 ▸ · at 18:05 ▸ · at 20:52 ▸
When this applies
when deciding whether to add steps to any conversion funnel
The takeaway
Don't minimize friction universally — test and keep friction that helps the user (or you) understand fit, and strip friction that doesn't.
Why it works
Friction that collects information to personalize downstream experience (recommendations, life-cycle targeting, look-alike audiences) pays for itself in conversion and retention; friction that merely delays without adding understanding only costs conversion.
One level up
Optimize a funnel for information value per step, not step count.
Where it stops holding
This stops working when you can't or don't test rigorously — adding friction on intuition alone, without data, risks killing conversion for the wrong reasons.
Detail
rule: Add friction when it helps users understand why the product is for them and enables better downstream personalization; cut friction that doesn't add value.
threshold: Test before trusting the simple 'cut everything' intuition.
rationale: Masterclass's quiz-in-purchase-flow and Mercury's split-screen onboarding both increased conversion by adding tested friction.
In their words
I've just seen time and time again at every job I've been in in growth that adding friction and adding the right steps leads to higher conversion and higher funnel completion.
Mercury onboarding quality push
Source
Lenny's Podcast: Product | Career | Growth — Head of Growth (Anthropic): “Claude is growing itself at this point” | Amol Avasare (April 5, 2026) · at 18:05 ▸
When this applies
when the first experience of your product is neglected relative to the quality bar you hold for the rest of the product
The takeaway
Sometimes the highest-leverage growth move is to temporarily abandon metric optimization and just fix experiential quality end-to-end.
Why it works
First impressions anchor trust and completion; investing purely in experiential quality (not metric-chasing) on the first touchpoint removes hidden conversion-killing friction that metrics-driven iteration hadn't caught.
One level up
Quality and growth are not separate levers — in high-stakes first touchpoints, quality IS the growth lever.
Where it stops holding
This stops working when the onboarding flow is already high quality, in which case a quality-focused quarter yields diminishing returns versus targeted experiments.
Detail
situation: Mercury's onboarding flow for regulated banking entities was complex and had never received the same quality investment as the rest of the product.
decision: The growth team dedicated an entire quarter to fixing pure experiential quality in onboarding, explicitly setting aside metrics as the goal.
outcome: Significant uplift in onboarding start-to-completion; Amol calls it the most impactful growth quarter of his career until joining Anthropic.
lesson: Quality drives growth, especially at first-touch moments, even when pursued without a direct metrics target.
In their words
forget metrics, forget growth, forget everything else. As the growth team on conversion, we're going to spend a whole quarter fixing quality in this flow.
Freedom through constraints — Anthropic's early resource-poor focus
Source
Lenny's Podcast: Product | Career | Growth — Head of Growth (Anthropic): “Claude is growing itself at this point” | Amol Avasare (April 5, 2026) · at 1:06:28 ▸ · at 1:09:44 ▸
When this applies
when an organization or individual faces severe resource constraints relative to competitors
The takeaway
Severe constraints can produce clarity and focus that abundance doesn't — a narrow bet made from necessity can outperform diffuse effort made from abundance.
Why it works
A hard constraint eliminates excess optionality, forcing a narrow, decisive focus (coding + B2B) that a well-resourced competitor spreading bets thin would not naturally arrive at, and that narrow focus also created a compounding feedback loop (better coding tools accelerate AI research itself).
One level up
Treat constraints as a forcing function for focus rather than only as a limitation to escape.
Where it stops holding
This stops working when the narrow focus chosen under constraint turns out to be the wrong market — Amol himself notes 'who knows' how it would have gone if Anthropic had launched its chatbot before ChatGPT and been pulled toward consumer instead.
Detail
situation: Anthropic lacked Meta/Google's distribution and cash, and OpenAI's first-mover advantage; it had chosen not to launch its own chatbot before ChatGPT for safety reasons.
decision: Leadership committed deeply and narrowly to AI coding and B2B from ~2021, partly from strategic conviction (coding accelerates research) and partly from necessity.
outcome: Anthropic grew from $1B to $19B+ ARR in 14 months, becoming (per the hosts) the fastest-growing company in history.
lesson: A deep, narrow focus chosen under resource constraint, reinforced by a self-accelerating feedback loop (better coding → faster research → better models), can outcompete better-funded generalists.
In their words
there's like this broader principle I have just around life of like the freedom through constraints that when you have a bunch of constraints applied on you, whether that's in personal life or at work